FACEBOOK HOSTS AND THIRD PARTY POSTS: DEFAMATION IN THE INTERNET AGE

HUGO DOBSON

A dissertation submitted in partial fulfilment of the degree of Bachelor of Laws (with Honours) at the University of Otago

October 2015

ii      

ACKNOWLEDGEMENTS

To Marcelo Rodriguez Ferrere, for his enthusiasm, insight and guidance as supervisor of this dissertation; and

to my parents, for their love and support.

iii      

FACEBOOK HOSTS AND THIRD PARTY POSTS: DEFAMATION IN THE INTERNET AGE

HUGO DOBSON

TABLE OF CONTENTS INTRODUCTION .................................................................................................................... 1

PART I: THE DEFAMATION TORT .................................................................................. 3

1.1 Origins of the defamation tort .......................................................................................... 3 1.2 Ideological underpinnings: freedom of expression and protection of reputation ............. 3

1.3 The tort ............................................................................................................................. 4

1.3.1 Publication generally ................................................................................................. 5

1.3.2 The first line of authority: publication by omission .................................................. 6

1.3.3 The second line of authority: publication by action (mitigated by the innocent dissemination defence) ....................................................................................................... 6

1.4 The Defamation Act 1992 ................................................................................................ 7

1.4.1 Codification of the innocent dissemination defence .................................................. 8

1.5 Conclusion ........................................................................................................................ 9

PART II: THE INTERNET .................................................................................................. 10 2.1 Social Media ................................................................................................................... 10

2.1.1 Facebook: Profiles, Pages, Groups, and Events ...................................................... 11

2.1.2 Facebook: the role of the host .................................................................................. 13

2.2 Differences in internet communication .......................................................................... 14

2.2.1 Transmissibility ....................................................................................................... 15

2.2.2 Accessibility ............................................................................................................ 15

2.2.3 Anonymity ............................................................................................................... 16

2.3 Conclusion ...................................................................................................................... 17

PART III: THE PROCEEDINGS ........................................................................................ 18 3.1 Facts ............................................................................................................................... 18

3.2 Wishart v Murray in the High Court .............................................................................. 19

3.2.1 A two-limbed test for third party liability on Facebook .......................................... 19

3.2.2 The High Court’s confused application of both lines of authority .......................... 20

3.2.3 The High Court’s use of analogies .......................................................................... 20

3.3 Murray v Wishart in the Court of Appeal ...................................................................... 21

iv      

3.3.1 The Court of Appeal’s use of analogies .................................................................. 22

3.3.2 The Court of Appeal’s concerns with the ought to know test ................................. 23

3.4 Conclusion ...................................................................................................................... 26

PART IV: THE COMMON LAW ....................................................................................... 27

4.1 Oriental Press: publication by action ............................................................................. 27 4.2 Tamiz: the mixed approach ............................................................................................ 28

4.3 Options for determining liability .................................................................................... 29

4.3.1 ‘Mixed’ liability ....................................................................................................... 30

4.3.2 Liability by action .................................................................................................... 30

4.3.3 Liability by omission ............................................................................................... 34

4.4 The use of analogies in legal reasoning ......................................................................... 38

4.5 Conclusion ...................................................................................................................... 39

PART V: LEGISLATIVE INTERVENTION ..................................................................... 41 5.1 ‘Regulatability’ of the internet ....................................................................................... 41

5.2 Regulatory approaches to new technology ..................................................................... 42

5.3 The Harmful Digital Communications Act: safe harbour provisions ............................ 44

5.3.1 Remove first; ask questions later ............................................................................. 46

5.3.2 Conclusions on the HDCA ...................................................................................... 47

5.4 The equivalent scheme in the United Kingdom ............................................................. 47

5.5 Conclusion ...................................................................................................................... 50

CONCLUSION ...................................................................................................................... 51 BIBLIOGRAPHY .................................................................................................................. 53  

 

1      

INTRODUCTION

 Chris and Cru Kahui were three-month old twins who died in Auckland in 2006. They died

from non-accidental injuries.1 Their father, Chris Kahui, was tried and acquitted of their

murder. During his trial he suggested the twins’ mother, Macsyna King, had inflicted the fatal

injuries.2 This suggestion “retained some currency in the public arena”.3 Ian Wishart is an

investigative journalist. Following the inquest into the twins’ deaths, he collaborated with

Macsyna King to produce a ‘tell-all’ book called Breaking Silence.4 In response to the book’s

impending release, Christopher Murray established a Facebook page called ‘Boycott the

Macsyna King Book’ in June 2011.5 The page peaked at around 50,000 ‘likes’ before being

taken down by Mr Murray.6 Mr Wishart filed proceedings in the High Court against Mr

Murray (along with his wife, Kerri Murray), alleging, inter alia, that comments from third

parties posted on the Facebook page defamed him.7 Mr Murray sought to strike out Mr

Wishart’s proceedings, and the High Court largely dismissed that application in March 2013.8

Mr Murray appealed against that decision, and the Court of Appeal released its judgment in

September 2014.9

These proceedings bring into focus the challenges involved in applying the defamation tort to

the internet, a medium which it was patently never intended to cover.10 My dissertation

explores these challenges. The specific legal question at the heart of my dissertation is this:

what is the appropriate legal test for determining when the host of a Facebook page will

become the publisher of a third party’s defamatory statement posted on that page? While

interesting in itself, this rather narrow legal question begets larger questions about the future

of the defamation tort, and about the best approach to regulating the internet. My dissertation

attempts to address both classes of question.

                                                                                                               1 Wishart v Murray [2013] NZHC 540 at [1]. 2 At [1]. 3 At [1]. 4 At [2]. 5 At [2]. 6 At [14], [80]. 7 At [3]; the claim relating to posts made by third parties on the Facebook page was in fact the fourth of four alleged causes of action. 8 Wishart v Murray, above n 1. 9 Murray v Wishart [2014] NZCA 461. 10 See 1.1 below.

2      

It is in five parts. Part I sets out the defamation tort in general terms. In particular, I

emphasise that there are two methods by which a publisher can become liable: by action, and

by omission. Part II considers the nature of the internet as a medium of communication, and

how its innovations have created challenges for the defamation tort. It also examines how

Facebook pages operate in practice. Part III critically examines the High Court and Court of

Appeal’s rulings in Murray v Wishart. I focus especially on the decision to apply a test of

liability by omission. Part IV situates the proceedings within the broader common law

landscape, and examines the alternative presented by a case that applied a test of liability by

action. I conclude that none of the approaches used in the common law are perfect, making

legislative intervention more attractive. Part V discusses the merits of legislative intervention,

both from a general perspective, and in terms of the Harmful Digital Communications Act

2015. Ultimately I conclude that while it is not without fault, the Harmful Digital

Communications Act makes a welcome complement to the defamation tort for striking the

appropriate balance between protecting the rights to reputation and freedom of expression

online.  

3      

PART I: THE DEFAMATION TORT

 This part introduces the tort of defamation. I discuss, in turn, the origins of defamation law,

its ideological underpinnings, how it functions in practice, and the Defamation Act 1992.

1.1 Origins of the defamation tort

The law of defamation is essentially aimed at the prohibition of the publication of injurious

false statements.11 Such an action can be traced, in general terms, back to the Bible, the

Mosaic Code, and the Talmud.12 However, the outlines of the modern law of defamation

emerged during the 16th and 17th centuries.13 It is essentially a product of the advent of the

printing press,14 the common law courts’ desire to expand their jurisdiction,15 and the Star

Chamber’s efforts to eradicate duelling.16 The English law of defamation thus has, in the

words of one commentator, a “variegated history”, from which New Zealand’s law of

defamation has developed.17

1.2 Ideological underpinnings: freedom of expression and protection of reputation

The defamation tort aims to protect the individual’s reputation without intruding too far into

the right to freedom of expression. The nature of defamation law means that these two values

are inevitably opposed.18 Thus the key challenge is in shaping the boundaries of the tort such

that the conflict between these two interests is minimised. This challenge is especially present

when the law is confronted with novel situations, like those thrown up by new technology.

Freedom of expression is an essential foundation of a democratic society.19 In particular, it is

conducive to the healthy operation of democratic institutions.20 It is also fundamental to the

                                                                                                               11 Hill v Church of Scientology of Toronto [1995] 2 SCR 1130 at [140]. 12 Peter F Carter-Ruck, Richard Walker, and Harvey NA Starte Carter-Ruck on Libel and Slander (4th ed, Butterworths, London, 1992) at 17. 13 William Holdsworth History of English Law (7th ed, Methuen, London, 1972) at 378. 14 At 378. 15 At 378. 16 Hill v Church of Scientology of Toronto, above n 11, at [116]. 17 David Rolph Reputation, Celebrity and Defamation Law (Ashgate, Aldershot, 2008) at 58. 18 Matthew Collins The Law of Defamation and the Internet (3rd ed, Oxford University Press, Oxford, 2010) at 4. 19 Handyside v United Kingdom (1976) 1 EHRR 737 at 754, cited with approval in Living World Distributors Ltd v Human Rights Action Group Inc (Wellington) [2000] 3 NZLR 570; (2000) 6 HRNZ 28 (CA) at 44 per Richardson P. 20 Hill v Church of Scientology of Toronto, above n 11, at [104].

4      

dignity of the individual, as its inclusion in the New Zealand Bill of Rights Act 1990

demonstrates.21

However, as Elias CJ noted in Brooker v Police, the right to freedom of expression is not

absolute.22 The defamation tort is one example of where the right is seen to be circumscribed

by a legitimate competing interest. Indeed, defamatory expression sits at odds with the values

underlying the right to freedom of expression. False and injurious forms of expression are

“inimical to the search for truth”, and “harmful to the interests of a free and democratic

society”.23

Unlike the right to freedom of expression, there is no right protecting an individual’s

reputation in the New Zealand Bill of Rights Act. However there are compelling reasons for

the existence of a tort that addresses the effect of injurious false statements. A good

reputation is closely related to an individual’s sense of self-worth and dignity.24 Once a

reputation has been tarnished by defamation, it will “seldom regain its former lustre”.25

1.3 The tort

Defamation is the communication of a false statement about a person that harms that person’s

reputation. The defamation tort aims to vindicate the plaintiff’s reputation and provide

compensation for the grief and distress caused by injury to reputation.26 It does so by

providing remedies in the form of damages, as well as injunctions and retractions.

A successful defamation action requires the plaintiff to make out that the defendant has

communicated a defamatory statement to someone other than the plaintiff.27 A statement will

be defamatory if, in the famous formulation of Lord Atkin, it tends to lower the individual

                                                                                                               21 Section 14: Freedom of expression: Everyone has the right to freedom of expression, including the freedom to seek, receive, and impart information and opinions of any kind in any form. 22 Brooker v Police [2007] NZSC 30 at [3]-[4]; Chief Justice Elias noted that the s 14 right draws on art 19 of the International Covenant on Civil and Political Rights. Article 19(3)(a) provides that the right to freedom of expression is subject to certain restrictions, including when necessary to protect the “rights or reputations of others”. 23 Hill v Church of Scientology of Toronto, above n 11, at [109]. 24 At [110]. 25 At [111]. 26 Television New Zealand Ltd v Keith [1994] 2 NZLR 84 (CA) at 86. 27 Patrick Milmo and WVH Rogers Gatley on Libel and Slander (10th ed, Sweet & Maxwell, London, 2004) at [1.3].

5      

concerned “in the estimation of right-thinking members of society generally”.28 A statement

may also be defamatory if it tends to cause the person to be shunned or avoided,29 or tends to

cause the person to be exposed to hatred, contempt or ridicule,30 or if it is a false statement

about a person to his or her discredit.31 A statement is not just something written or oral, but

may take the form of visual images, sounds, gestures, or other means of signifying

meaning.32

1.3.1 Publication generally

A person will be liable for a defamatory statement when they are deemed to have published

the statement.

Generally, ‘publication’ occurs when a person intentionally or negligently takes part in, or

authorises, the communication of defamatory material to someone other than the plaintiff.33

There is no publication, and therefore no cause of action, if the material is communicated

only to the plaintiff.34 In an internet context, publication is usually deemed to occur when

material is downloaded or accessed (rather than uploaded) from a web server.35 The

defamation tort deems each communication of defamatory material to be a separate

publication.36 Defamatory material on a webpage is thus usually ‘published’ for the purposes

of defamation each time it receives a ‘hit’ from an internet user.37 Moreover, every person

who knowingly takes part in the communication of defamatory material is prima facie

liable.38 The original author is plainly liable, but equally is anyone who takes part in, or

authorises, the publication of another’s defamatory material.39 This wide ambit of potential

                                                                                                               28 Sim v Stretch (1936) 52 TLR 669 (HL) at 671. 29 Youssoupoff v Metro-Goldwyn-Mayer Pictures Ltd (1934) 50 TLR 581 (CA) at 587. 30 Parmiter v Coupland (1840) 6 M&W 105 at 108. 31 Scott v Sampson (1882) 8 QBD 491 at 503. 32 Monson v Tussauds Ltd [1894] 1 QB 671 at 692. 33 Collins, above n 18, at 67. 34 Brian Neill, Richard Rampton, Heather Rogers, Timothy Atkinson and Aidan Eardley Duncan and Neill on Defamation (3rd ed, LexisNexis, London, 2009) at [8.01]; Pullman v Walter Hill & Co [1891] 1 QB 524 at 527. 35 Godfrey v Demon Internet Ltd [2001] QB 201 at 208; Dow Jones & Co Inc v Gutnick [2002] HCA 56; (2002) 210 CLR 575 at [137]. 36 Neill, Rampton, Rogers, Atkinson and Eardley, above n 34, at [8.07]. 37 At [8.07]. This rule has been mitigated in the United Kingdom by the ‘single publication rule’ in s 8 of the Defamation Act 2013 (UK), which deems that subsequent publications of substantially the same statement will not be considered new publications, such that the one year limitation period does not reset every time a website is accessed. 38 Collins, above n 18, at 67. 39 At 101.

6      

liability is subject to the defence of innocent dissemination, which I discuss below.40 The

person who publishes the defamatory material need not know that what they are publishing

contains defamatory material, rather it is the simple act of communication that gives rise to

liability.41

1.3.2 The first line of authority: publication by omission

Liability may arise both by action and omission. The leading authority on publication by

omission is a case called Byrne v Deane.42 In this case the proprietors of a golf club were held

responsible for the publication of an allegedly defamatory notice posted, without permission,

on a wall of the golf club. The King’s Bench held that because the proprietors were aware of

the notice and could have easily removed it, but did not do so, they were liable as publishers.

In the words of Greene LJ, the test for liability by omission was:

“…having regard to all the facts of the case is the proper inference that by not removing the defamatory matter the defendant really made himself responsible for its continued presence in the place where it had been put?”43

Thus publication by omission occurs when the defendant somehow ratifies or acquiesces in

the original publication. For the sake of clarity and consistency I will refer to these principles,

and their attendant line of authority, as ‘the Byrne v Deane line’. Subsequent cases have

applied the Byrne v Deane line to various contexts,44 and more recently, and with mixed

success, courts have applied the Byrne v Deane line to publication on the internet.45

 

1.3.3 The second line of authority: publication by action (mitigated by the innocent

dissemination defence)

The Byrne v Deane line of authority stands in contrast to the line of authority that deals with

publication by positive action. Instead of focusing on the conditions by which omission can

lead to liability, this line concerns situations where the defendant is actively involved in the

                                                                                                               40 At 1.3.3. 41 Milmo and Rogers, above n 27, at [6.17]. 42 Byrne v Deane [1937] 1 KB 818. 43 At 838. 44 For example Isabelle Hellar v Joe Bianco (1952) 244 P 2d 757; Urbanchich v Drummoyne Municipal Council (1991) Aust Tort Reports 81-127 (NSW SC). 45 For example Godfrey v Demon Internet Ltd, above n 35; Davison v Habeeb [2011] EWHC 3031 (QB); Tamiz v Google Inc [2012] EWHC 449 (QB); [2013] EWCA Civ 68.

7      

act of publication. Every person who takes part in the chain of publication is held prima facie

liable as a publisher.46

However, the breadth of activity captured by the traditional publication rule is vast,47 and the

strictness of this rule is mitigated by the ‘innocent dissemination defence’. The leading case

in this line is Emmens v Pottle,48 which, along with Vizetelly v Mudie’s Select Library,49 gave

rise to what eventually became known as the defence of ‘innocent dissemination’.50

Traditionally, the common law grants protection to a person who published defamatory

material, but was not that material’s author, printer, or first publisher, and had only taken a

“subordinate part in disseminating it”.51 Lord Justice Romer’s original statement of the

defence required that the person seeking to rely on the defence: (1) did not know the work

contained the libel52 complained of; (2) did not know the work was of a character likely to

contain a libel; and (3) that such lack of knowledge was not due to any negligence by that

person.53 The onus to make out the defence lies with the defendant, and the question of

liability is one for the jury.54

For the sake of clarity and consistency I will refer to this line of authority, where publication

is prima facie deemed to have occurred, but the innocent dissemination defence may be made

out, as the ‘Emmens v Pottle line’.

1.4 The Defamation Act 1992

The Defamation Act 1992 partially codified the law of defamation in New Zealand.

                                                                                                               46 Collins, above n 18, at 101. 47 Crookes v Newton [2011] 3 SCR 269 at [18]. 48 Emmens v Pottle (1885) 16 QBD 354. 49 Vizetelly v Mudie’s Select Library Limited [1900] 2 QB 170. 50 Robert Ribeiro “Defamation on the Internet” (paper presented to Obligations VII Conference, Hong Kong, July 2014) at [13]. 51 Milmo and Rogers, above n 27, at [6.18], from Vizetelly v Mudie’s Select Library Limited, above n 49, at 180. 52 ‘Libel’ is the traditional common law term for defamation in written form. Historically, the common law treated libel slightly differently to slander (oral forms of defamation), with libel actionable per se but slander requiring proof of actual damage. This distinction was collapsed by s 4 of the Defamation Act 1992. 53 Summarised from Vizetelly v Mudie’s Select Library Limited, above n 49, at 180. 54 At 180.

8      

Its substance is much older than its year of enactment suggests. It is the result of the 1977

commission of a government-appointed committee to investigate the state of defamation law

in New Zealand.55 In 1997, Sir Geoffrey Palmer explained the difficult passage of the

Defamation Bill:

“The Muldoon Government would not do anything on the [committee’s report]; they did not touch it. Put it on a shelf and forgot about it but, when the Labour Party got in, as Minister of Justice I was keen to advance the reform. We finally got a Defamation Bill in but it was not easy because Members of Parliament do not like reforming the law of defamation. They think that the law of defamation is a useful thing with which to beat the media, and many of them regarded it as quite inappropriate that the law should be changed. I had a lot of trouble with that Bill as a result. It was not passed while I was in Parliament, it was not passed until 1992.”56

It would reasonably be thought that legislation enacted in 1992 would have contemplated the

internet, and the possible issues it might create. But the substance of the Act actually dates

from the late 1970s, and thus did not – indeed, could not – have done so.

1.4.1 Codification of the innocent dissemination defence

Section 21 of the Defamation Act codified the innocent dissemination defence in New

Zealand:

21 Innocent dissemination In any proceedings for defamation against any person who has published the matter that is the subject of the proceedings solely in the capacity of, or as the employee or agent of, a processor or a distributor, it is a defence if that person alleges and proves— (a) that that person did not know that the matter contained the material that is alleged to be

defamatory; and (b) that that person did not know that the matter was of a character likely to contain

material of a defamatory nature; and (c) that that person's lack of knowledge was not due to any negligence on that person's

part. … 2 Interpretation …

distributor includes— (a) a bookseller; and (b) a librarian … processor means a person who prints or reproduces, or plays a role in printing or reproducing, any matter …

                                                                                                               55 Ursula Cheer “Defamation in New Zealand and Its Effects on the Media – Self-Censorship or Occupational Hazard?” (2006) NZ L Rev 467 at 468, n 3. 56 Geoffrey Palmer Constitutional Conversations: Geoffrey Palmer talks to Kim Hill on National Radio (Victoria University Press, Wellington, 2002) at 175.

9      

An important feature of the defence is that it only applies to processors and distributors (or

their agents and employees). Whether this includes internet media is unclear. In other

jurisdictions, where the defence has not been codified, it is typically available to a person

who played a “subordinate part in disseminating” the defamatory material in the ordinary

course of business.57 This more generic formulation would seem to more readily cover

internet media. Nevertheless it could still apply to internet media in New Zealand.58

1.5 Conclusion

This part has outlined the defamation tort. Importantly, there are two ways publication can

occur – through action, and through omission – both of which have attendant lines of

authority. This dynamic begs two questions in relation to third party publication on

Facebook. First, does the Facebook host become liable by action or omission? Second, if by

omission, what standard of knowledge will found publication – actual knowledge, or

something less?

The New Zealand High Court and Court of Appeal had to answer these questions in the

Murray v Wishart proceedings, and I examine their handling of them in Part III. Part IV will

address these questions more directly. Now, however, it is useful to examine the medium in

question. This is the task of Part II.

                                                                                                               57 Vizetelly v Mudie’s Select Library Limited, above n 49, at 180. 58 See 3.3.2(e) below for a fuller discussion of this point in light of the Court of Appeal’s position in Murray v Wishart.

10      

PART II: THE INTERNET

 The differences between the internet and traditional print media are significant. This part

considers how the internet has changed how society communicates, and the challenges these

changes pose to the defamation tort. I begin by examining how the particular medium at issue

in this dissertation – the Facebook page – operates in practice. I then compare the internet to

traditional print media, focusing particularly on the increased transmissibility and

accessibility of online material, and the ease with which it can be disseminated anonymously.

2.1 Social Media

The term ‘social media’ refers to a group of internet-based applications that harness the

power of the internet to encourage the creation and exchange of user-generated content.59

Traditionally, mass media adopted a ‘one-to-many’ form of communication.60 This means

that information originates from a single source, and is published to many recipients. The

original publisher has editorial control over the content, and the ability to choose what is and

is not published.

Social media platforms are distinctive because they are a type of ‘many-to-many’

communication: from the people, to the people.61 The operators of social media platforms do

not necessarily broadcast themselves, but instead facilitate the speech of others.62 Their

business models depend on providing widespread and democratised access to media and

encouraging participation from users.63

                                                                                                               59 For a fuller discussion of the term see Andreas M Kaplan and Michael Haenlein “Users of the world, unite! The challenges and opportunities of social media” (2010) 53 Business Horizons 59 at 61. 60 Ribeiro, above n 50, at [5]. 61 Jack Balkin “Media Access: A Question of Design” (paper presented to the Global Constitutionalism Seminar, Yale Law School, 2010) at 1813. 62 At 1813. 63 At 1813.

11      

2.1.1 Facebook: Profiles, Pages, Groups, and Events

Facebook is a social media platform launched in 2004. Since then its use has grown

significantly. In June 2015, Facebook reported an average of 968 million daily active users

and 1.49 billion monthly active users.64

Facebook users begin by creating a ‘Profile’. The Profile may include a profile picture,

contact information, and other personal information. Users can then interact with other users

by ‘friending’, sending private or public messages, and sharing content.65

The Profile essentially operates as an online persona for a user.66 But it is not the only type of

Facebook page. Once a user has created a Profile, he or she can then use it to create a ‘Page’,

‘Group’, or ‘Event’.

A ‘Page’ is usually based around an interest, hobby, organisation, brand or cause. Other users

can elect to ‘like’ the Page to receive updates from it. There is a huge range of Pages on

Facebook, from the predictable – Portuguese footballer Cristiano Ronaldo has the largest

individual following with 106.7 million likes – to the obscure – the ‘I Will Go Slightly Out of

My Way To Step On A Crunchy-Looking Leaf’ Page has attracted 1.7 million likes.67

Pages are the most commercialised type of Facebook page and therefore have the most

complicated administrative structure. There are five different types of role for people who

manage Pages.68 ‘Admins’, ‘Editors’, and ‘Moderators’ can all delete posts and ban users

from accessing the Page.69

                                                                                                               64 Facebook Newsroom <http://newsroom.fb.com/company-info> accessed 5 October 2015. 65 Facebook Help Center: How to Post & Share <www.facebook.com/help/333140160100643/> accessed 9 September 2015. 66 Facebook Help Center: Pages Basic <www.facebook.com/help/217671661585622> accessed 9 September 2015. 67 <www.facebook.com/Cristiano>; <www.facebook.com/pages/I-Will-Go-Slightly-Out-of-My-Way-To-Step-On-A-Crunchy-Looking-Leaf/128579817570> accessed 2 October 2015. 68 Facebook Help Center: Page Roles <www.facebook.com/help/323502271070625/> accessed 9 September 2015. 69 Facebook Help Center: Page Roles, above n 68, accessed 9 September 2015; see Fig. 1.

12      

A ‘Group’ is a dedicated space where members can share content and message other

members.70 Some Groups are ‘secret’, which means only members invited into the Group by

the Group’s creator can see that it exists or access it. Other Groups are ‘closed’, meaning

non-members can see that the group exists but only access some of its content. Others still are

‘open’, meaning anyone can view and post in the Group.71 The creator of the Group is

automatically made an ‘Admin’, with the power to remove members, edit the Group’s

settings, and make other members Admins.72

Page roles Admin Editor Moderator Advertiser Analyst Manage Page roles and settings ü Edit the Page and add apps ü ü Create and delete posts as the Page ü ü Send messages as the Page ü ü ü Respond to and delete comments and posts to the Page

ü ü ü

Remove and ban people from the Page ü ü ü Create ads ü ü ü ü View insights ü ü ü ü ü See who published as the Page ü ü ü ü ü

Figure 1: Five different types of Page role73

An ‘Event’ is a page usually dedicated to some real world event such as a social gathering or

promotion. Like Groups, Events have different levels of privacy: some can be accessed only

by those who have been invited by a Host, while others are accessible to anyone on the

internet.74 In the same way that Groups can have multiple ‘Admins’, Events can have

multiple ‘Hosts’.75

                                                                                                               70 Facebook Help Center: Group Basics <www.facebook.com/help/162866443847527/> accessed 9 September 2015. 71 Facebook Help Center: Group Basics, above n 70, accessed 9 September 2015. 72 Facebook Help Center: Group Admin Basics <www.facebook.com/help/418065968237061/> accessed 9 September 2015. 73 Facebook Help Center: Page Roles, above n 68, accessed 9 September 2015. 74 Facebook Help Center: Events Privacy <www.facebook.com/help/216355421820757/> accessed 9 September 2015. 75 Facebook Help Center: Creating & Editing Events <www.facebook.com/help/131325477007622/> accessed 9 September 2015.

13      

2.1.2 Facebook: the role of the host

The different types of Facebook page have subtly different rules around who can view, post

and moderate content. The Admin of a Page is essentially the same as the Admin of a Group

and the Host of an Event. However, the position becomes more complicated when there is

more than one Admin or Host, or when there is one Admin and a team of Moderators.76

This dissertation contemplates the situation where there is a single host, as was the case in

Murray v Wishart.77 For the sake of simplicity, I use the label ‘host’ to refer to the individual

who created the page and continues to have ultimate control over its content. Likewise I use

the term ‘page’ as an umbrella term covering Profiles, Pages, Groups and Events. I use the

term ‘third party liability’ to refer to the Facebook host’s liability for posts that were made to

his or her page by someone else.

When the host creates a page, the content posted on the page is stored on a server maintained

by Facebook.78 As such the host of a page does not actually host the content per se, rather he

or she gains certain abilities in relation to controlling and moderating the page. The host can

delete other users’ posts, prevent posts containing certain words from being posted, and

remove or ban users from the page.79

Who can view the page’s content depends on what type of page it is. All types of page have a

range of privacy settings that the host can change. Profiles are usually private, with most

information available only to ‘Friends’ of the host, although it is possible to make

information publicly available.80 Pages (with a capital ‘P’) are generally intended to be

public, so are typically accessible by everyone, although the host can set age and country

restrictions, and can proactively moderate content by blocking words or setting a profanity

filter.81 Groups and Events have varying degrees of privacy that the host can choose. They

                                                                                                               76 See Fig. 1 above at 2.1.1. 77 Murray v Wishart, above n 9. 78 Collins, above n 18, at 29. 79 Facebook Help Center: Managing a Page: Moderation <www.facebook.com/help/329858980428740/> accessed 11 August 2015. 80 Facebook Help Center: How to Post & Share, above n 65, accessed 9 September 2015. 81 Facebook Help Center: Managing a Page <www.facebook.com/help/255700674532721/> accessed 9 September 2015.

14      

may require invitation to the Group or Event to access content, or the content may be publicly

available.82

One of the great complications around the intersection of the defamation tort with Facebook

is the range of different types of Facebook page. Pages differ immensely in both their

purpose and reach. While their fundamental technical aspects are largely the same, they can

differ vastly in tenor and function.

This means that when it comes to regulating Facebook hosts, someone who checks their

Facebook Profile once a month, and someone who is commercially employed to moderate a

fan Page, will in theory be subject to the same rule. The differences in the type and volume of

material that pages attract means the moderating responsibilities of a host can vary hugely.

Thus in deciding when a Facebook host becomes a publisher of someone else’s statement

posted on his or her page, it is desirable that the legal test takes into account the existence of

this spectrum.

2.2 Differences in internet communication

The modern law of defamation grew out of the advent of the printing press,83 and the tort of

defamation is traditionally geared towards written and oral communications. But the internet

has radically changed – and continues to change – the way in which humans interact and

communicate. As Matthew Collins puts it, the internet “has already displaced, or is in the

process of displacing, to a greater or lesser degree, every form of written communication”.84

It seems then that the tort of defamation is being applied to forms of communication for

which it was patently not conceived. I turn now to discuss the novelties in these new forms of

communication, and the challenges they create for the defamation tort.

                                                                                                               82 Facebook Help Center: Group Admin Basics, above n 72; Facebook Help Center: Events Privacy, above n 74, accessed 9 September 2015. 83 See 1.1 above. 84 Collins, above n 18, at 33.

15      

2.2.1 Transmissibility

Publishing something online is much easier than publishing something in print. The Law

Commission recognised this in 1999.85 Instead of having to print and physically distribute

something, a Facebook account can be used to reach hundreds of people in seconds. The

internet has reduced the cost of publication to virtually zero. In the words of Mark Tushnet:

“In a world in which libelous statements can be distributed only through a newspaper, there will be fewer such statements than in a world where they can be distributed via the Internet, simply because it is cheaper to log on to the Internet than to purchase a newspaper.”86

Online media also make it much easier to re-publish. If a message or image is sufficiently

interesting or scandalous its republication can accelerate rapidly, in a phenomenon well

known as ‘going viral’. This means the potential for defamation to spread is greatly amplified

in online media, both in terms of the material’s reach and the speed with which it is

republished. In the words of one commentator, the “extraordinary capacity of the Internet to

replicate almost endlessly any defamatory message lends credence to the notion that ‘the

truth rarely catches up with a lie’ ”.87

This feature of the internet has also been noted in a judicial context. In Cairns v Modi Lord

Judge CJ, in the course of justifying a large award of damages, recognised that

“as a consequence of modern technology and communications systems, any such stories will have the capacity to “go viral” more widely and more quickly than ever before. Indeed it is obvious today, with the ready availability of the world wide web and of social networking sites, the scale of this problem has been immeasurably enhanced.”88

2.2.2 Accessibility

A second defining feature of material posted on the internet is that it can remain accessible

indefinitely, and can be accessed by people all over the world. As Cardozo J observed about

the difference between oral and written material:

“What gives the sting to writing is its permanence in form. The spoken word dissolves, but the written one abides and perpetuates the scandal.”89

                                                                                                               85 Law Commission Electronic Commerce Part 2: A basic legal framework (NZLC: R58, 1999) at [262]. 86 Mark Tushnet “Internet Exceptionalism: An Overview From General Constitutional Law” (2015) 56 Wm and Mary L Rev 1637 at 1655. 87 Lyrissa Barnett Lidsky “Silencing John Doe: Defamation & Discourse in Cyberspace” (2000) 49 Duke LJ 855 at 863-64. 88 Cairns v Modi [2013] 1 WLR 1015; [2012] EWCA Civ 1382 at [27]. 89 Ostrowe v Lee 256 NY 36 (NY Ct App 1931) at 39.

16      

The internet gives the written word an additional level of permanence. While oral gossip

tarnishes a reputation, its effect fades over time. But the internet makes gossip a permanent

reputational stain: “one that never fades”.90

The problem of online material being accessible indefinitely is compounded by the

development of increasingly intelligent search engines.91 Instead of having to manually trawl

through old newspapers or records, search engines can sift through vast amounts of data in

seconds. Jack Balkin argues that the internet operating in this way evokes a “Nietzschean

image of persistent memory… of eternal return”.92

It should be noted that this may not continue to be true. A recent decision of the Court of

Justice of the European Union affirmed the existence in the EU of a so-called ‘right to be

forgotten’ – specifically, the right to have personal data deleted from search engines on

request.93 This is a developing area of law, but there is nothing to indicate that a similar right

will be recognised in New Zealand any time soon.

It seems that – the possible development of ‘right to be forgotten’ jurisprudence aside –

searchable digital records of our lives, accessible from anywhere and by anyone, are simply a

feature of modern life.

2.2.3 Anonymity

The internet also makes it much easier to publish a statement anonymously.94 Internet users

are able to conceal their identity relatively easily behind “an impenetrable cloak of

anonymity”.95 If an individual does not fear being held responsible for publishing something

                                                                                                               90 Daniel J Solove The Future of Reputation: Gossip, Rumor, and Privacy on the Internet (Yale University Press, New Haven, 2007) at 33. 91 Ryan J Turner “Internet Defamation Law and Publication by Omission: A Multi-Jurisdictional Analysis” (2014) 37 UNSWLJ 34 at 56. 92 Balkin, above n 61, at 1813. Nietzsche posited the idea of “eternal return”: that the universe has been recurring, and will continue to recur an infinite number of times: see Friedrich Nietzsche The Gay Science (Cambridge University Press, Cambridge, 2001). 93 Case C-131/12 Google Spain SL and Google Inc v Agencia Española de Protección de Datos (AEPD) and Mario Costeja González OJ 2014/C 212/04; for commentary on the impact of this decision see Eleni Frantziou “Further Developments in the Right to be Forgotten: The European Court of Justice’s Judgment in Case C-131/12, Google Spain, SL, Google Inc v Agencia Española de Protección de Datos” (2014) 14 HRL Rev 761. 94 Law Commission, above n 85, at [262]; David Rolph “Defamation by social media” (2013) 117 Precedent 16 at 18. 95 Collins, above n 18, at 21.

17      

that could be defamatory, he or she will be more willing to publish it. Whereas in the past

publishing something anonymously took a great deal of effort, the internet has made it

virtually effortless.96

This means that the Facebook host – or indeed anyone who has the ability to moderate an

online forum – may have to contend with anonymous users. Although Facebook requires the

registration of some personal details on the Profile page, there is nothing to ensure these

details are accurate. And while tracking someone through their internet protocol (‘IP’)

address is sometimes possible, it is complex.97 Accordingly, in some situations the Facebook

host stands to become responsible for comments of third parties where the originator of the

post cannot be traced.

2.3 Conclusion

If the defamation tort is to be applied to the internet, the fundamental challenge is working

out how the tort can continue to effectively protect reputation without crushing the great

potential of the internet as a medium for public expression. But publication on the internet is

significantly different to publication via traditional print media. Compared to print media,

publication on the internet is characterised by the ease with which material can be published,

the increase in the material’s longevity, and the ability to publish anonymously.

This set of challenging conditions amounts to a new paradigm of publication for the

defamation tort. Indeed, if the changes wrought by the internet can rightly be called

“revolutionary”,98 then it must be considered whether instead of a clumsy attempt to apply

the principles of defamation to internet media, it could be more desirable to implement

technology-specific rules governing malicious falsehoods online.

I will return to this question in due course, but turn now to examine the Murray v Wishart

proceedings.  

                                                                                                               96 Tushnet, above n 86, at 1645. 97 H Brian Holland “In Defense of Online Intermediary Immunity: Facilitating Communities of Modified Exceptionalism” (2007) 56 U Kan L Rev 369 at 393: “Advances in technology are making it increasingly possible to locate and identify bad actors online, such that online anonymity is difficult to maintain”. 98 Solove, above n 90, at 17.

18      

PART III: THE PROCEEDINGS

 This part focuses on the Murray v Wishart proceedings.99 The proceedings bring into sharp

relief the conflict of principles outlined in Part I and the technological realities discussed in

Part II.

3.1 Facts

Chris and Cru Kahui were three-month old twins who died in Auckland in 2006. They died

from non-accidental injuries, and their deaths generated a significant amount of public

interest.100 Their father, Chris Kahui, was tried and acquitted of their murder. During his trial

he suggested the twins’ mother, Macsyna King, had inflicted the fatal injuries.101 This

suggestion “retained some currency in the public arena”.102

The plaintiff, Ian Wishart, was an investigative journalist. Following the inquest, he

collaborated with Macsyna King to produce a ‘tell-all’ book called Breaking Silence.103

In response to the book’s impending release, the first defendant, Christopher Murray,

established a Facebook page called ‘Boycott the Macsyna King Book’ in late June 2011 (‘the

Facebook page’).104 The page peaked at around 50,000 ‘likes’ before being taken down by

Mr Murray on or around 13 August 2011.105 It was a ‘Page’, meaning that it was publicly

accessible.106

Mr Wishart filed proceedings against Mr Murray, alleging, inter alia, that comments from

third parties posted on the Facebook page defamed him.107 Mr Murray sought to strike out Mr

Wishart’s proceedings.108

                                                                                                               99 Murray v Wishart, above n 9; Wishart v Murray, above n 1. 100 Wishart v Murray, above n 1, at [1]. 101 At [1]. 102 At [1]. 103 At [2]. 104 At [2]. 105 At [14], [80]. 106 See 2.1.1 above for an explanation of the different types of Facebook page and their respective characteristics. 107 At [3]; the claim relating to posts made by third parties on the Facebook page was in fact the fourth of four alleged causes of action.

19      

3.2 Wishart v Murray in the High Court

It is appropriate to repeat the questions posed at the end of Part I: first, does the Facebook

host become liable by action or omission? Second, if by omission, what standard of

knowledge will found publication – actual knowledge, or something less?

Justice Courtney found that Mr Murray could became liable for publication by omission

under the Byrne v Deane line of authority.109 She held that either actual or constructive

knowledge could found publication.

3.2.1 A two-limbed test for third party liability on Facebook

Justice Courtney held that the host of a Facebook page is regarded as a publisher of posts

made by third parties to that page in two situations. The first is if he or she knows of the

defamatory statement and fails to remove it within a reasonable time, in circumstances that

give rise to an inference that he or she is taking responsibility for the statement (the ‘actual

knowledge’ test).110 The second situation is where he or she does not know of the defamatory

statement but ought, in the circumstances, to know that third parties are making posts that are

likely to be defamatory (the ‘ought to know’ test).111

Figure 2: High Court held ought to know and actual knowledge tests apply

                                                                                                                                                                                                                                                                                                                                                       108 At [4]: the strike out application was made on two grounds: that the statements in issue were not capable of bearing the pleaded meanings, and that the statement of claim is prolix and oppressive and contains material that is scandalous and/or irrelevant. 109 See 1.3.2 above. 110 At [117]. 111 At [117].

Early liability

Late liability

Strict liability

Ought to know

Actual knowledge

20      

3.2.2 The High Court’s confused application of both lines of authority

Justice Courtney arrived at this outcome following an examination of case law from several

jurisdictions and from both lines of authority. While Courtney J held the appropriate test for

publication was to be found in the Byrne v Deane line (publication by omission), she

emphasised the ratio in Emmens v Pottle.112 In doing so she conflated the two lines of

authority that deal with publication by action and publication by omission, treating Emmens v

Pottle as if it were part of the Byrne v Deane line of authority.113 As the Court of Appeal

recognised on appeal:

“we think the better view is that [Emmens v Pottle] is about innocent dissemination and therefore is not authority for the proposition that a person may be found to have published a defamatory statement on the ground that they ought to have known of its existence”.114

Interestingly, some of Courtney J’s analysis indicates she was alive to the distinction. She

employed the Emmens v Pottle ratio correctly in analysing one of the few relevant New

Zealand authorities, Sadiq v Baycorp.115 In Sadiq the plaintiff complained about defamatory

statements concerning his creditworthiness on a debt collector’s website which had been

placed online by the website’s previous owner. Justice Courtney held that the defendant

should have been considered a publisher in the first instance, with the real issue then being

the availability of the innocent dissemination defence.116 Unfortunately, Justice Courtney’s

appropriate use of the Emmens v Pottle authority in analysing Sadiq was obscured by her

treatment of the authority elsewhere.

3.2.3 The High Court’s use of analogies  

Justice Courtney approved of the ‘noticeboard analogy’ that was made in Davison v

Habeeb117 and approved by the English Court of Appeal in the Tamiz appeal.118 This analogy

compared the intermediary (in both Davison and Tamiz it was the Google blogging service

‘Blogger.com’) to a

                                                                                                               112 At [85]. 113 At [85]; see 1.3.2 and 1.3.3 above for explanations of the two lines of authority. 114 Murray v Wishart, above n 9, at [96]. 115 Sadiq v Baycorp (NZ) Ltd HC Auckland CIV-2007-404-6421, 31 March 2008; Wishart v Murray, above n 1, at [89]-[90]. 116 At [90]. 117 Davison v Habeeb, above n 45, at [38]. 118 Wishart v Murray, above n 1, at [116]; Tamiz v Google Inc, above n 45.

21      

“giant noticeboard … in the sense that [Google] provides the noticeboard for users to post their notices on, and it can take the notices down (like the club secretary in Byrne v Deane) if they are pointed out to it”.119

The comparison between Google (or the Facebook host) and the club secretary in Byrne v

Deane is undermined by the fact that, in Byrne v Deane, rather than soliciting notices, it was

against the club rules to post notices without the proprietors’ approval.120 This distinction,

between solicitation and trespass, is an important one, as it counts towards the solicitor (or

Facebook host) being considered a publisher at first instance, instead of applying the Byrne v

Deane line of authority relating to publication by omission.121

This is nearly but not quite the conclusion Courtney J came to. She found that those who host

Facebook pages “are not passive instruments”.122 However, this did not justify treating them

as prima facie publishers (with the possibility of making out a defence). Instead, Courtney J

saw this fact as relevant to the standard of knowledge requirement, leading her to conclude

that constructive knowledge was sufficient to found publication by omission.

Justice Courtney found there was at least a tenable case on the facts that Mr Murray was a

publisher of the third party posts.123 She also considered that it was open to him to avail

himself of the innocent dissemination defence.124

3.3 Murray v Wishart in the Court of Appeal

Mr Murray appealed Courtney J’s ruling on several grounds. Important for present purposes

is the Court of Appeal’s treatment of the appeal as it related to the legal test for determining

whether the host of a Facebook page is the publisher of statements posted on it by third

parties. President O’Regan and Ellen France J, delivering the judgment for the Court,125 held

the appropriate legal test solely consisted of the actual knowledge test and not, contrary to the

High Court’s decision, the actual knowledge test as well as the ought to know test.126 In so

                                                                                                               119 Davison v Habeeb, above n 45, at [38]. 120 Byrne v Deane, above n 42, at 837. 121 See 4.1 below for more on the implications of this distinction. 122 Wishart v Murray, above n 1, at [117]. 123 At [122]. 124 At [122]. 125 Justice French concurred. 126 Murray v Wishart, above n 9, at [144].

22      

doing they, like Courtney J, applied Byrne v Deane principles relating to publication by

omission.

Figure 3: Court of Appeal held only actual knowledge test applies

The Court of Appeal saw “the authorities evaluated by Courtney J and cited to us as

providing limited guidance”.127 Instead the Court emphasised the importance of focusing on

the particular factual situation, noting that the different roles taken by the various alleged

publishers in each of the authorities gave rise to different outcomes.128 The Court ultimately

found there was “no precedent requiring us to adopt the ought to know test”,129 and

abandoned the test based on several “concerns”.130 In applying the law to the facts, the Court

discussed the merits of several analogies from the authorities.131

3.3.1 The Court of Appeal’s use of analogies

Following an acknowledgment that “analogies are only analogies”, the Court of Appeal

examined three analogies that could be made to the Facebook host.132 The three analogies

were:

(a) the posting of a notice on a notice board (or a wall on which notices can be affixed) without the knowledge of the owner of the notice board/wall; [the Byrne v Deane situation]

(b) the writing of a defamatory statement on a wall of a building without the knowledge of the building owner; [the graffiti situation] and/or

(c) a defamatory comment made at a public meeting without the prior knowledge or subsequent endorsement or adoption by the organiser of the meeting.133 [emphasis added]

                                                                                                               127 At [125]. 128 At [125]. 129 At [136]. 130 At [136]ff. 131 At [125]ff. 132 At [126]. 133 At [126].

Early liability

Late liability

Strict liability

Ought to know

Actual knowledge

23      

The Court considered that the notice board analogy was at least somewhat appropriate,

although it noted that in Byrne v Deane the board had been posted on in trespass, whereas a

Facebook host solicits posts.134 The Court found a second difference between the notice

board and the Facebook host was that the post to the notice board was “truly anonymous”,

whereas many of the people posting on the Facebook page were identifiable.135 While this

may have been true in the case of Mr Murray’s Facebook page, it will not always be so. Fake

or anonymous Facebook profiles are easily created.136

The Court did not consider the graffiti analogy as particularly helpful, essentially for the

same trespass point as the notice board analogy.137

The Court considered the public meeting analogy most useful.138 Its attraction appeared to lie

in illustrating a situation where “even if a person incites defamation, he or she will not

necessarily be liable for defamatory statements made by others”.139 The Court acknowledged

that the analogy is not perfect, as statements at a meeting would be oral and therefore

ephemeral, unlike the written comments on a Facebook page.140 This is a significant

difference, but the Court of Appeal did not appear to accord it much weight.141

3.3.2 The Court of Appeal’s concerns with the ought to know test

The Court listed five concerns with the ought to know test.142

(a) A Facebook host who does not know about a defamatory post is worse off

The first concern was that the ought to know test puts the Facebook host who does not know

about a defamatory post to his or her page in a worse position than a host who does know

about it.143 The reason given for this was that the Facebook host who ought to have known

about the defamatory post will never have an opportunity to avoid liability by the timely

                                                                                                               134 At [129]; see 4.1 below. 135 At [130]. 136 See 2.2.3 above. 137 At [131]. 138 At [132]. 139 At [134]. 140 At [133]. 141 See 2.2.2 above; also see 4.4 below for a discussion of the use of analogies in legal reasoning generally. 142 At [136]ff. 143 At [137].

24      

removal of the post. But under the actual knowledge test, the host is allowed a reasonable

time to remove the post.

In finding that a host has constructive knowledge, a court is assessing when it is reasonable to

deem that the host ought to have known that defamatory comments were likely to be made to

the page. It seems at least theoretically possible to also give the host who has constructive

knowledge the benefit of a reasonable time to remove the comments.

(b) Strict liability is contrary to the intentional nature of the tort

The second concern was that the ought to know test would make the host liable on a strict

liability basis, or on the basis of negligence, and to do so would be contrary to the “well-

understood nature of the tort of defamation as an intentional tort”.144 This concern was not

explained in depth, and would have benefited from greater analysis.

The case cited as authority for the proposition that defamation is an intentional tort, Bell-

Booth Group Ltd v Attorney-General, dealt with an attempt to argue for a new duty of care to

not injure someone’s reputation by true statements.145 The Court of Appeal in this case

rejected all of the appellant’s arguments, and so while the case does seem to stand against the

bleeding of negligence into defamation law, it was decided in the context of a duty not to

injure by true statements.

The sense in which the traditional rule requires publication to be ‘intentional’ is generally

directed at situations where the defendant who authored the defamation does not intend its

publication to a third person, but where publication takes place without negligence on his or

her part.146 An example, given by Lord Esher MR, is where “the writer of a letter locks it up

in his own desk, and a thief comes and breaks open the desk and takes away the letter and

makes its contents known”.147 The requirement of ‘intention’ deems that in situations like this

the writer is not a publisher. Thus it is difficult to see how the ‘intentional nature’ of the tort

                                                                                                               144 At [140]. 145 Bell-Booth Group Ltd v Attorney-General [1989] 3 NZLR 148 (CA) at 156. 146 Oriental Press Group Ltd v Fevaworks Solutions Ltd [2013] HKCFA 47 at [21]. 147 Pullman v Walter Hill & Co, above n 34, at 527.

25      

would be at odds with the imposition of strict liability or liability through negligence. Indeed,

this is what the traditional publication rule achieves.148

(c) Conflict with the right to freedom of expression

The third concern was a bald assertion that the ought to know test conflicts with the right to

freedom of expression affirmed in s 14 of the New Zealand Bill of Rights Act 1990.149 The

Court of Appeal asserted that the ought to know test unduly favours the interests of a person

whose reputation is damaged over the right to freedom of expression.150 It did not explain

why.

(d) Uncertainty in application

The fourth concern was that the ought to know test is uncertain in its application.151 This is a

legitimate concern. Given the ubiquity of Facebook in the modern age, it is certainly

desirable that the law provides clear boundaries for liability. This is the strongest argument in

favour of an actual knowledge test.

(e) The innocent dissemination defence would not apply to a Facebook host

The fifth concern was that the innocent dissemination defence would not apply to a Facebook

host.152 The Court of Appeal held that the defence, now statutorily provided for in s 21 of the

Defamation Act 1992,153 was unlikely to be available to a Facebook host.

The Court of Appeal considered the language of s 21 “aimed at old media” and “inapplicable

to internet publishers”.154 However the terms ‘processor’ and ‘distributor’ are at least

somewhat open ended in their definitions.155 Distributor merely includes booksellers and

                                                                                                               148 The rule is often described as imposing strict liability: see, for example, Panday v Gordon [2006] 1 AC 427 at 436 per Lord Nicholls of Birkenhead: “The basic position at which the common law holds this balance is to impose strict liability for defamatory statements in the absence of justification”. That publication can occur through negligence is also relatively uncontroversial: see, for example, Collins, above n 18, at 67; Joachim Dietrich “Clarifying the meaning of ‘publication’ of defamatory matter in the age of the internet” (2013) 18 Media and Arts Law Review 88 at 89. 149 Murray v Wishart, above n 9, at [141]. 150 At [141]. 151 At [142]. 152 At [143]. 153 See 1.4.1 above. 154 At [143]. 155 See 1.3.3 above for full definitions.

26      

librarians.156 Processors are not only those who print, but also those who reproduce, “any

matter”.157 It seems at least open to a court to include internet publishers under s 21.

In support of the unavailability of the innocent dissemination defence, the Court of Appeal

cited the Law Commission’s recommendation that the defence be statutorily amended to

extend to ISPs (Internet Service Providers, such as Spark or Vodafone).158 The Court implied

that the Law Commission’s position was that the defence would not apply to internet

publishers. In fact, the Law Commission is less clear:

“While, in ECom 1, we indicated that an ISP would probably fall within the definition of “processor” and “distributor”, on reflection we tend to the view that the law should be amended to remove any residual doubt.”159

It should also be noted that ISPs have a different function to Facebook hosts in terms of the

role they play in publication. The Law Commission’s position on the innocent dissemination

defence thus offers, at best, weak support for the Court of Appeal’s assessment of its

unavailability. It remains open to the Court to take the view that the innocent dissemination

defence could, in the appropriate circumstances, apply to Facebook hosts. A similar view was

reached by the Hong Kong Court of Final Appeal in Oriental Press in relation to the host of

an internet discussion forum.160 However, the Court of Appeal has demonstrated its

reluctance to take this position with its ruling in Murray v Wishart.161

3.4 Conclusion

Both the High Court and the Court of Appeal applied Byrne v Deane publication by omission

principles to assess the Facebook host’s liability for third party content. They differed in

terms of the standard of knowledge required to found such liability, with the Court of Appeal

ruling that only actual knowledge will suffice. Part IV will situate these positions within their

broader common law context, beginning with an exploration of a case that found that

publication by action was the appropriate approach to assessing third party liability for an

online content host.162  

                                                                                                               156 Defamation Act 1992, s 2. 157 Section 2. 158 At [143]. 159 Law Commission, above n 85, at [269]. 160 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146; see 4.1 below. 161 At [143]. 162 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146.

27      

PART IV: THE COMMON LAW

 This part surveys the legal landscape around the Murray v Wishart proceedings. It begins by

analysing two alternatives to the approaches taken in those proceedings. The first is found in

a decision that rejected the Byrne v Deane approach and instead adopted publication by

action principles (the Emmens v Pottle line) to assess third party liability online.163 The

second is found in a decision that applied both lines of authority.164 This part then evaluates

the respective merits of both alternatives presented in these cases, as well as the approaches

taken by the High Court and Court of Appeal.165 It ends with a brief note on the limitations of

using analogies in legal reasoning.

4.1 Oriental Press: publication by action

In Oriental Press, the Hong Kong Court of Final Appeal held that the providers of an online

discussion forum – the ‘Hong Kong Golden Forum’ – were publishers of defamatory material

posted on the forum by third parties but were entitled to rely on the defence of innocent

dissemination.166 The leading judgment, delivered by Ribeiro PJ, found that the relevant

question in determining the forum providers’ liability for third party material was not whether

the forum providers had become publishers of the material, but whether they could rely on

the innocent dissemination defence.167

Permanent Justice Ribeiro emphasised that there are important differences between the

principles that make up the two lines of authority.168 The innocent dissemination defence

provides relief to subordinate publishers who knowingly participate in the process of

disseminating defamatory material.169 In contrast, occupiers in the Byrne v Deane line of

authority did not assist or participate in the originator’s publication.170 The Byrne v Deane

line of authority instead deals with “the conditions which must be satisfied before the

occupier can be regarded as having turned himself into a publisher”.171

                                                                                                               163 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146. 164 Tamiz v Google Inc, above n 45. 165 Murray v Wishart, above n 9; Wishart v Murray, above n 1. 166 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146. 167 At [52]. 168 At [46]; see 1.3.2 and 1.3.3 above for the two lines of authority. 169 At [47]. 170 At [48]. 171 At [48].

28      

As Ribeiro PJ noted (extra-judicially), the Byrne v Deane authorities tend to concern acts of

trespass.172 In the original case, the offending verse was posted without the golf club

proprietors’ permission.173 Isabelle Hellar v Joe Bianco concerned defamatory graffiti on the

wall of the men’s bathroom in a bar.174 Urbanchich v Drummoyne Municipal Council

concerned defamatory posters glued onto several bus shelters.175

Because of this feature of trespass, Ribeiro PJ held in Oriental Press that it was inappropriate

to apply the Byrne v Deane principles to internet platform providers:

“The provider of a discussion forum is in a wholly different position from that of the occupier of premises who is not in the business of publishing or facilitating publication at all, but who has had imposed on him the defamatory act of a trespasser.”176

In contrast, the forum providers played an “active role” in encouraging and facilitating the

posts made by members of their forum, making them clear participants in the publication

“from the outset”.177 This distinction, between trespass and solicitation, is central to the

Oriental Press ruling. The forum providers’ solicitation of posts was the key reason why the

Court found that instead of falling under the Byrne v Deane line of authority, the forum

providers were prima facie publishers under the Emmens v Pottle line.178

4.2 Tamiz: the mixed approach

In Tamiz the English Court of Appeal found that Google was potentially liable in defamation

for failing to take down allegedly defamatory content. In assessing Google’s liability, the

Court of Appeal applied both lines of authority relating to publication: publication by action

principles for the period before receiving notice of complaint about the content, and

publication by omission principles for after that period.179

The plaintiff complained that a blogger had defamed him on Blogger.com, a blogging

platform provided by Google.180 Google forwarded his complaints to the blogger, who

                                                                                                               172 Ribeiro, above n 50, at [30]. 173 Byrne v Deane, above n 42, at 837. 174 Isabelle Hellar v Joe Bianco, above n 44. 175 Urbanchich v Drummoyne Municipal Council, above n 44. 176 At [50]. 177 At [51]-[52]. 178 At [51]-[52]. 179 Tamiz v Google Inc, above n 45, at [25]-[36]. 180 At [2].

29      

removed the offending content three and a half months after the original complaint was laid

(but only five weeks after Google was deemed to have received notice of complaint).181 Mr

Tamiz sued Google in respect of the period before the content was removed.182

The Court of Appeal held that Google was not a primary publisher because it did not have

any prior knowledge of, or control over, the content of the blog.183 Lord Justice Richards,

delivering the judgment for the Court, doubted whether Google was even a subordinate

publisher, but nevertheless held that before receiving Mr Tamiz’s complaint, Google had a

defence because it did not know and would not, by exercising reasonable care, have known

that the publication was defamatory.184 Though the words ‘innocent dissemination’ were not

used, it is clear that this is what Richards LJ was referring to, as he cited both Emmens v

Pottle and Vizetelly v Mudie’s Select Library.185

However, in deciding the nature of post-notification liability, Richards LJ applied Byrne v

Deane principles.186 The five week period that elapsed between notification and removal was

held to be “sufficiently long to leave room for an inference adverse to Google Inc on Byrne v

Deane principles”.187

Notwithstanding his finding that Google was a publisher on this basis, Richards LJ held the

defamation in question was trivial and therefore dismissed the appeal.188

4.3 Options for determining liability

Oriental Press and Tamiz illustrate that there are (at least) three distinct methods of

determining the third party liability of a Facebook host. The more straightforward approach

taken by the cases so far is to apply the Byrne v Deane (publication by omission) line of

authority, as was adopted in the Murray v Wishart proceedings by both the High Court and                                                                                                                181 At [2] and [9]; while Mr Tamiz first notified his complaint on 28 or 29 April 2011, Google was deemed to have received notice of complaint sometime in July, making the relevant period between notification and removal “over five weeks” (at [35]). 182 At [3]. 183 At [25]. 184 At [26]. 185 At [26]: Emmens v Pottle, above n 48; Vizetelly v Mudie’s Select Library Limited, above n 49; see 1.3.3 above. 186 At [27]-[36]. 187 At [35]. 188 At [50].

30      

Court of Appeal. The second option, taken in Oriental Press by the Hong Kong Court of

Final Appeal, is to apply the Emmens v Pottle (publication by action) line of authority, which

makes the defendant prima facie liable but with the benefit of the defence of innocent

dissemination. The third option is to take the Tamiz approach, and apply Emmens v Pottle

principles to the period before notification, and Byrne v Deane principles to the period after.

4.3.1 ‘Mixed’ liability

As Ribeiro PJ (who delivered the leading judgment in Oriental Press) noted extra-judicially,

the English Court of Appeal in Tamiz appears to have considered it possible to adopt both the

innocent dissemination defence, and the Byrne v Deane approach.189 The substance of the

innocent dissemination defence – whether the defendant ought to have known about the

defamatory statement and took reasonable care – is similar to that of the Byrne v Deane

approach – whether the defendant deliberately or negligently failed to remove defamatory

material after being put on notice. However, the principles underlying each approach do not

sit comfortably together in concert: “They address different issues; apply different standards

in judging the defendant’s conduct, and impose different burdens of proof.”190

It is important that the courts adopt a coherent and consistent approach to defamation on the

internet.191 The mixed approach taken in Tamiz undermines this goal, and risks engendering

confusion by introducing a complicated double standard into the law.

4.3.2 Liability by action

The Oriental Press approach has some appeal. It imposes a strong standard of liability on

content hosts, with the onus of proof on the host to make out a defence.192 Making the

defendant prima facie liable in this way increases the standard of liability that he or she is

held to. Matthew Collins has contended that the practical differences stemming from this

reversal of onus would likely be insignificant.193 However, the principled distinction is

important. If the Facebook host can rightly be said to be a participant in the chain of

                                                                                                               189 Ribeiro, above n 50, at [31]. 190 Ribeiro, above n 50, at [31]. 191 Turner, above n 91, at 42. 192 Vizetelly v Mudie’s Select Library Limited, above n 49, at 180. 193 Collins, above n 18, at 117.

31      

publication, then under the traditional publication rule, this is the position that sits most

comfortably with the existing body of defamation law.

However, applying the Oriental Press approach to Facebook hosts is potentially problematic

for two reasons.

(a) Facebook hosts are arguably materially different to forum providers

There are significant differences between a Facebook host and the providers of an online

forum. Forum providers like those in Oriental Press play a highly active role:  

“…they designed the forum …; they laid down conditions for becoming a member …; they employed administrators whose job was to monitor discussions and to delete postings which broke the rules; and they derived income from advertisements placed on their website, a business model which obviously benefits from attracting as many users as possible to the forum.”194

Comparing forum providers to Facebook hosts is complicated by the different types of

Facebook page.195 Some pages are maintained, like the ‘Hong Kong Golden Forum’, for

commercial benefit.196 But others, like Mr Murray’s ‘Boycott the Macsyna King Book’ page,

are motivated by non-commercial public causes or interests. Others still do not aspire to

attract public exposure at all.

It makes sense to hold a commercial entity that derives income from the provision of a

service to a higher standard of liability than an individual who may run a blog or a Facebook

page on a casual basis. Commercial entities are more likely to have the resources necessary to

effectively moderate online fora. If they aspire to profit from providing a medium for public

discourse, it follows that they should internalise the cost of moderating that medium.

However, in general, Facebook hosts are much more hands-off than the forum providers in

Oriental Press. The degree of effort the forum providers expended in soliciting third party

content was central to Ribeiro PJ’s ruling in Oriental Press, as it meant the forum providers

could not say they were not part of the chain of publication. This factor is greatly reduced in

the case of a Facebook host. To deem the Facebook host part of the chain of publication on

the basis that he or she solicits posts to his or her page is potentially problematic. While the

                                                                                                               194 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [51]. 195 See 2.1.1 above. 196 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [51].

32      

solicitation is often present, it is far less obvious, and the Facebook host is typically far more

passive than the forum providers in Oriental Press.

This nuance was overlooked by Courtney J, who found in Wishart v Murray that it was

implicit in the decision in Tamiz that individuals who create and control blogs should be

regarded as essentially in the same position as Google (who provided the blogging

platform).197 This proposition is questionable, both in terms of whether it is legitimately

implicit in the Tamiz ruling, and in its factual accuracy generally.

Lord Justice Richards did rule that Google’s role in respect of Blogger blogs was not a purely

passive one.198 But the basis on which this could be extended to blog hosts (or Facebook

hosts) is unclear. Importantly, Google operates the Blogger service as a commercial

activity.199 This is true also of the operation of Facebook, and of the forum providers in

Oriental Press.

There is also a considerable difference between forum providers who have to monitor

thousands of threads, and most Facebook hosts who are usually responsible for one, or at

most a handful, of pages. 200 Although especially popular Facebook pages may attract a huge

volume of material, even such that a legitimate comparison to the ‘Hong Kong Golden

Forum’ might be made, this type of page represents a small fraction of the pages on

Facebook.201

Thus while Courtney J is likely correct that Facebook hosts “are not passive instruments”,202

nor can they rightly be said to be directly analogous to commercial entities who provide

social media platforms.

                                                                                                               197 Wishart v Murray, above n 1, at [114]; Tamiz v Google Inc, above n 45; see 4.2 above for analysis of Tamiz and 3.2 above for Courtney J’s ruling. 198 Tamiz v Google Inc, above n 45, at [23]. 199 At [24]. 200 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [92]. 201 In 2012, when Facebook had 526 million active daily users, there were only 42 million Pages with 10 or more likes: Brian Hongiman “100 Fascinating Social Media Facts and Figures From 2012” <www.huffingtonpost.com/brian-honigman/100-fascinating-social-me_b_2185281.html> 202 Wishart v Murray, above n 1, at [117].

33      

(b) The innocent dissemination defence may not apply to Facebook hosts

It was a central feature of the Oriental Press decision that the Court considered that the

innocent dissemination defence was available to the forum providers.203 However it is unclear

whether the New Zealand courts consider it appropriate to apply the innocent dissemination

defence to internet media.

The Court of Appeal indicated in Murray v Wishart that the innocent dissemination defence

“would be difficult to apply” to a Facebook host.204 While I suggested above that it was open

to the Court to take the position that the defence could be available,205 following Murray v

Wishart this would require somewhat of a judicial about-turn.

The essential difficulty appears to arise from the decision to codify the defence in s 21 of the

Defamation Act.206 The Court of Appeal identified that the “language of the section and the

defined terms used in it” meant that it appeared inapplicable to internet publishers.207 The

common law defence left in generic terms appears more amenable to application to internet

media than the codified form of the defence in s 21. Notably, Hong Kong does not have a

codified defence and Ribeiro PJ’s analysis of the availability of the defence was based on its

common law form.208

The effect of the codification of the defence in New Zealand is ossification. The common law

is often lauded for its ability to adapt to reflect the progression of society and development of

technology.209 It is tempting then to lament the decision to codify the defence as short-

sighted. Whether Parliament’s intention in codifying the defence was to limit it to traditional

print forms of media is unclear, but regardless this appears to be what it has achieved.210

                                                                                                               203 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [89]. 204 Murray v Wishart, above n 9, at [143]. 205 At 3.3.2(e). 206 See 1.4.1 above. 207 Murray v Wishart, above n 9, at [143]. 208 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [30]. 209 For example Home Office v Harman [1983] 1 AC 280 at 320 per Lord Roskill: “New situations regularly arise in the practice of the law which require previously held and sometimes generally accepted views to be reviewed and if necessary to be revised in the light of that new situation. Indeed, the evolution of the common law of this country to meet the changing needs of contemporary society and its adaptability to change owes much to judicial acceptance of this philosophy.” 210 In fact, when the Defamation Act was drafted Parliament likely did not even know what the internet was, much less consider the impact of codifying the defence on internet media. The Defamation Act’s year of enactment is 1992, but its substance is actually much older: see 1.4 above.

34      

If the Court of Appeal is correct that the innocent dissemination defence is unavailable to

internet media, the result would be a broad imposition of liability under the traditional

publication rule. A Facebook host would effectively become strictly liable for anything

posted on his or her page. This does not alone constitute sufficient grounds for rejecting the

Emmens v Pottle publication by action line of authority, but negative consequences could

result from such a broad imposition of liability.

Ryan J Turner argues that litigation against internet hosts, especially internet media

corporations, is likely to increase substantially.211 He cites Richard Posner as describing this

and last century as “an age when tort law is dominated by the search for the deep pocket”.212

If this is true, the cost of defamation litigation may simply become a business expense for

internet media. Related to this is the concern that an expansive test for liability risks deterring

companies, public institutions (including universities) and individuals from establishing

online discussion fora because of the risk of liability for defamatory material posted on that

forum.213 While arguably this is appropriate in the case of commercial entities that can

internalise the costs of moderation,214 it is important to consider the possible chilling effect

on freedom of expression. In seeking to protect reputation without unjustifiably intruding into

the right to freedom of expression, the defamation tort is inevitably engaged in a balancing

act – one which risks being upset if individuals and institutions are deterred from creating

online fora for public discourse.215

Thus while there is certainly merit in the Oriental Press approach, especially in terms of its

principled coherence with existing defamation law, it is also seriously problematic in several

respects in the New Zealand context.  

4.3.3 Liability by omission

It seems then that the Byrne v Deane principles must provide the legal framework for

determining liability.216 In examining the merits of applying the Byrne v Deane line, it is first

                                                                                                               211 Turner, above n 91, at 61. 212 Turner, above n 91, at 205. 213 Turner, above n 91, at 61. 214 See 4.3.2(a) above. 215 See 1.2 above. 216 Murray v Wishart, above n 9; Wishart v Murray, above n 1; see 1.3.2 above.

35      

necessary to address Ribeiro PJ’s contention in Oriental Press that it only applies to

situations of trespass. The content of the knowledge requirement must also be considered,

and the more fundamental question of whether it is valid to compare the Byrne v Deane

situation to the case of a Facebook host.

(a) The Byrne v Deane line may only apply to situations of trespass

Implicit in Ribeiro PJ’s decision in Oriental Press is that it would be inappropriate to apply

the Byrne v Deane line to situations not involving trespass.217 However his characterisation of

the Byrne v Deane line as necessarily involving acts of trespass is open to debate. Turner

argues that such a requirement is “at odds” with the approach of the publication by omission

line of authority, which he contends has not been constrained by such a requirement.218 He

suggests that Byrne v Deane is better read as:

“involving an act of publication contrary to guidelines or rules set down for publication [no posting on the walls of the club without consent], as opposed to an act of publication constituting a trespass”.219

As explored above, the Byrne v Deane line of authority does tend to involve acts of

trespass.220 However this is not always the case. Notably, several of the cases that apply

Byrne v Deane principles to publication online do not involve trespass.221 Of course this does

not settle the question of whether trespass was – or should be – part of the Byrne v Deane

ratio. The issue of trespass is really about whether it makes sense to characterise the person in

question as part of the chain of publication. Where the original publication was made in

trespass (the classic example being graffiti),222 this counts against deeming the proprietor (or

Facebook host) a publisher. However, where the publication is not technically a trespass, but

still, for example (as in Turner’s characterisation of Byrne v Deane above) in breach of a

particular guideline or rule, this fact still counts against deeming the proprietor a publisher,

because it tends to reduce the blameworthiness of the proprietor for the act of publication. It

is a false binary to say that anything posted in trespass falls outside the publication rule, and

anything short of trespass falls within.

                                                                                                               217 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [50]. 218 Turner, above n 91, at 50. 219 Turner, above n 91, at 51. 220 At 4.1. 221 See, for example, Godfrey v Demon Internet Ltd, above n 35; Davison v Habeeb, above n 45; Tamiz v Google Inc, above n 45. 222 See 3.3.1 above.

36      

While Turner is probably correct that the Byrne v Deane line is not restricted to situations of

trespass, the circumstances around publication, including whether the original publication

was in breach of any rules (or laws) must still be an important consideration in an assessment

of third party liability. The characterisation of the publication in Byrne v Deane as in breach

of club rules instead of a trespass does little to reduce the force of the fundamental point that

it is unfair to hold the proprietors liable for someone else’s illegal (or prohibited) act.

Thus the Byrne v Deane line should not be rejected in the context of Facebook hosts simply

because someone posting to a Facebook page does not do so in trespass.

(b) The content of the knowledge requirement

Interestingly, the Hong Kong Court of Final Appeal briefly discussed Courtney J’s ruling223

in the Oriental Press case (the appeal not having been heard at that time).224 Permanent

Justice Ribeiro referred to the case in the context of discussing what conditions are needed to

satisfy the standard of reasonable care for the innocent dissemination defence.225 Wishart v

Murray was considered to highlight the “potential need for particular care in certain cases”

when individuals “mounted campaigns against targeted persons”.226 Most interestingly for

present purposes, Ribeiro PJ held:

“Unlike the hosts of a discussion forum involving many thousands of discussion threads, the creators of such a blog or Facebook page have a relatively narrow horizon to monitor and, given that their express purpose is to mount a campaign critical of others, they might reasonably be expected to exercise particular care in respect of potentially defamatory comments posted at their invitation.”227 [emphasis added]

Permanent Justice Ribeiro held that when someone has created a blog or Facebook page

expressly for the purpose of criticising someone, this could elevate what constitutes

reasonable care in the circumstances. This is an intuitively appealing point: such a page is

clearly more likely to attract defamatory content.

Here an important consequence to the distinction between the actual knowledge and

constructive knowledge tests is brought into focus.228 Taking into account the purpose of the

                                                                                                               223 Wishart v Murray, above n 1. 224 Oriental Press Group Ltd v Fevaworks Solutions Ltd, above n 146, at [92]. 225 At [92]. 226 At [91]-[99]. 227 At [92]. 228 See 3.2 and 3.3 above.

37      

page is possible under the publication by action line (through the innocent dissemination

defence, which is the context in which Ribeiro PJ discussed Wishart v Murray), and also

under Courtney J’s ‘ought to know’ test.229 However, this consideration would not be

relevant under the Court of Appeal’s ‘actual knowledge’ test, which is concerned simply with

whether or not the Facebook host knew the defamatory content had been posted.230 The

actual knowledge test sacrifices flexibility for certainty. It seems the Court of Appeal valued

the latter over the former.231

This is probably justifiable. Certainty is important in the law, and arguably especially so in

the case of the Facebook host given the near ubiquity of Facebook in New Zealand.232 The

potential for defamation on Facebook is truly vast, so it is desirable to have a simple rule

governing when the Facebook host will become liable for a third party’s statement. An actual

knowledge test would be easier for a layperson to understand than the more opaque ought to

know test, or the approach of prima facie liability mitigated by the innocent dissemination

defence.

(c) Facebook hosts are arguably materially different to golf club proprietors

However, if the Byrne v Deane principles are going to be applied, it is important to examine

the validity of equating Facebook hosts with golf club proprietors.

A Facebook page is fundamentally unlike a golf club notice board because it is an internet

medium. I have explored above at some length the paradigm shift that publication online

represents.233 While a golf club notice board (or bathroom graffiti) has a relatively restricted

potential audience, the capacity for defamation to spread on Facebook is virtually boundless.

Moreover, the Facebook host – albeit to varying extents, but certainly in the case of those

                                                                                                               229 Wishart v Murray, above n 1, at [117]; see 3.2.1 above. 230 Murray v Wishart, above n 9, at [144]; see 3.3 above. 231 At [142]. 232 According to one source New Zealand in fact has the “number one usage of social networks in the world”: Jo Brothers “A Social Media Breakfast with Teddy Goff and Ross Dawson” The Huffington Post (online ed, 3 April 2013) <www.huffingtonpost.com/jo-brothers/a-social-media-breakfast-_b_2703179.html>. Another source puts the number of New Zealanders aged 20-24 on Facebook at nearly 90%: <www.firstdigital.co.nz/blog/2014/09/10/facebook-demographics-new-zealand-age-distribution-gender/> accessed 3 October 2015. Also see 2.1.1 above. 233 At 2.2.

38      

who are promoting a cause, such as Mr Murray234 – solicits and encourages others to post to

the page. In contrast, the golf club proprietors prohibited posting to the notice board without

permission.235 Arguably then the key principle in Byrne v Deane – that the defendant

becomes liable only after somehow ratifying or taking responsibility for the original

publication – is inappropriate under this different set of conditions. Where there is such a

greater potential for defamatory content to go viral, and the Facebook host has (at least to

some degree) solicited the defamatory content, a stricter test for liability may be warranted.

Thus while I have argued above that there are material differences between a Facebook host

and the providers of an online forum,236 so too are there significant differences between a

Facebook host and the proprietors of a golf club. This is probably inevitable, and reflects the

inherently imperfect nature of reasoning by analogy.237

4.4 The use of analogies in legal reasoning

Part of the difficulty in determining the correct legal test for third party publication on the

internet comes from the existing common law authorities dealing with substantially different

situations. In spite of this, the courts employed reasoning by analogy in an attempt to stretch

the existing principles to an online context. It is appropriate to comment briefly on the

limitations of reasoning by analogy.

On one view, analogical reasoning is simply part of the inherent nature of reasoning under

the common law. As Ribeiro PJ noted extra-judicially, “it is a hallmark of the common law

that the courts seek to adapt established principles to solve novel problems”.238 Indeed, some

theorists regard analogies as the cornerstone of common law reasoning.239

However, reasoning by analogy also has staunch critics. Richard Posner argues that analogies

have no real normative force, rather:

                                                                                                               234 Murray v Wishart, above n 9; see 3.1 above. 235 Byrne v Deane, above n 42, at 837. 236 See 4.3.2(a) above. 237 See 4.4 below. 238 Ribeiro, above n 50, at [31]; Ribeiro acknowledges in the same paragraph that it “may seem odd that [the courts] have had recourse to centuries-old cases and have considered the possible value as precedents, of such matters as graffiti on lavatory walls and versified notices pinned onto golf club notice boards”. 239 See, for example, EH Levi An Introduction To Legal Reasoning (University of Chicago Press, Chicago, 1948) at 1-3.

39      

“Reasoning by analogy tends to obscure the policy grounds that determine the outcome of a case, because it directs the reader’s attention to the cases that are being compared with each other rather than to the policy considerations that connect or separate the cases. … There is no such thing as an “analogical argument” in any but a rhetorical sense; you need reasons to determine whether one case should be thought relevantly similar to another.”240

Perhaps the courts’ strained use of analogies reflects something in the nature of the

defamation tort. Joachim Dietrich argues that the courts’ difficulties in applying the

principles of publication to internet media really stem from “conceptual uncertainties” that

are pre-existing in the defamation tort.241 In his analysis, the inconsistent application of the

law in the cases concerning internet media reflects a fundamental instability in the underlying

legal doctrine, rather than the novel nature of publication on the internet.242

It is important to note that the courts are not blind to the limitations of reasoning by analogy.

The Court of Appeal in Murray v Wishart recognised that:

“Many of the decisions show an effort by the relevant Court to reach a conclusion by reference to an analogy in a non-internet circumstance. We agree that is a helpful form of reasoning, but it must be acknowledged that the analogies are only analogies, and there is room for debate about their appropriateness in particular cases.”243

The Court of Appeal was clearly aware that reliance on analogies could only prove so useful,

and its scepticism is healthy.244 However, it still ultimately applied principles from the Byrne

v Deane line of authority to decide the case.245

4.5 Conclusion

This part has evaluated various options presented by the common law for determining third

party liability for Facebook hosts. I have characterised the two key alternatives as liability

through action along Oriental Press lines, and liability through omission under Byrne v

                                                                                                               240 Richard A Posner “Book Review: Reasoning By Analogy” (2006) 91 Cornell L Rev 761 at 765 and 768. 241 Dietrich, above n 148, at 88. 242 Dietrich, above n 148, at 88. One of the uncertainties Dietrich explored is whether knowledge of the (defamatory) nature of the content is relevant to the question of publication: while the orthodox view is that knowledge of the nature of the content of the publication goes only to defences, some commentators have argued that where knowledge is absent, publication has not occurred – see, for example, Clerk & Lindsell on Torts (21st ed, Sweet & Maxwell, London, 2014) at [22-57]. 243 Murray v Wishart, above n 9, at [126]; and earlier, at [99]: “The analysis of the cases requires the Court to apply reasoning by strained analogy, because the old cases do not, of course, deal with publication on the internet. There is a question of the extent to which these analogies are helpful.” 244 Grant Lamond “Analogical Reasoning in the Common Law” (2014) 34 Oxford Journal of Legal Studies 567 at 587: “distant analogies have serious limitations, and a moderate scepticism about their use is not out of place”. 245 See 3.3 above.

40      

Deane principles. The Court of Appeal and High Court in the Murray v Wishart proceedings

took the latter approach, differing in their approaches to the requirements for knowledge.246

None of the approaches is perfect. Given the statutory construction of the innocent

dissemination defence in New Zealand, the Oriental Press approach would likely lead to an

undesirably strict standard of liability.247 It seems then that the Byrne v Deane line must

provide the appropriate legal framework. If so, the next question relates to the content of the

knowledge requirement. While a test based on constructive knowledge allows for a more

nuanced application of Byrne v Deane principles, a test based on actual knowledge provides a

more certain basis for assessing liability.

Ultimately the question of what is the appropriate legal test, and the difficulties therein, give

way to a larger question. If the attempt to apply the principles of defamation law to internet

media is inevitably going to be clumsy, could a better course be to implement technology-

specific rules governing malicious falsehoods online? Part V attempts to answer this

question.  

                                                                                                               246 See 3.2 and 3.3 above. 247 See 1.4.1 and 4.3.2(b) above.

41      

PART V: LEGISLATIVE INTERVENTION

 The alternative to applying the existing principles of defamation law to internet media is the

introduction of technology-specific rules governing defamation-type situations online. This

part considers such an alternative course. It begins by examining the regulation of the internet

in general terms. It then considers the recently enacted Harmful Digital Communications Act

2015, which affects the liability of an internet content host for third party content, and

compares this new statute to its equivalent in the United Kingdom.

5.1 ‘Regulatability’ of the internet

Early discussions about regulating the internet were dominated by optimism that the internet

was in fact ‘unregulatable’. This sentiment possibly reached its zenith in John Perry Barlow’s

declaration of the independence of cyberspace in 1996:

“Governments of the Industrial World. … You have no sovereignty where we gather. … I declare the global social space we are building to be naturally independent of the tyrannies you seek to impose on us. …”248

The idea that the internet could be conceptualised as a ‘state of being’ which had naturally

evolved apart and distinct from terrestrial space, was known as ‘exceptionalism’.249 In the

early 1990s, a vision of the online environment emerged which advocated and defended

internet exceptionalism, known as ‘cyberlibertarianism’.250 Cyberlibertarians envisioned

cyberspace as a distinct sphere, separate from physical space, which was, and should

continue to be, naturally self-regulating.251

This early period of cyberlibertarianism coincided with the enactment of the Communications

Decency Act 1996 in the United States.252 Section 230 of this Act provides exceptionalist253

                                                                                                               248 John Perry Barlow “A Declaration of the Independence of Cyberspace” <https://projects.eff.org/~barlow/Declaration-Final.html> accessed 29 September 2015. 249 Holland, above n 97, at 108. 250 Holland, above n 97, at 109. 251 Holland, above n 97, at 110. 252 47 USC § 230(1): “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” 253 Exceptionalist in the sense that it provides greater protection for online providers than offline publishers who publish the same content.

42      

protection from liability for providers of an “interactive computer service” who publish

information provided by others.254

Not all commentators shared the cyberlibertarians’ enthusiasm however. Frank Easterbrook

famously compared the ‘law of cyberspace’ to the “law of the Horse”:

“Lots of cases deal with sales of horses; others deal with people kicked by horses; still more deal with the licensing and racing of horses, or with the care veterinarians give to horses, or with prizes at horse shows. Any effort to collect these strands into a course on "The Law of the Horse" is doomed to be shallow and to miss unifying principles.”255

He argued that it is wrong to conceive of the law of cyberspace as somehow distinctive or

worthy of special treatment. To do so risked “multidisciplinary dilettantism”.256

Lawrence Lessig responded directly to Easterbrook’s criticism, arguing that the study of

regulation online can shed light on legal issues that are relevant to the law generally.257 He

cautioned that the ‘architecture’ of the internet – those structural features that constrain

behaviour – is a highly pervasive, and unobserved, agent of regulation.258 Lessig’s influence

on this area of scholarship was (and continues to be) substantial.259 The idea that the internet

is beyond the reach of real-space regulation is now much less popular; in Lessig’s own

words: “The confidence of the Internet exceptionalists has faded”.260

5.2 Regulatory approaches to new technology

The growing acceptance that the internet is amenable to real-space regulation has been

mirrored by the actual enactment of internet-specific legislation. However before I examine

two incidences of this, from New Zealand and the United Kingdom, it is worthwhile to

consider regulatory approaches to new technology generally.

                                                                                                               254 Holland, above n 97, at 102. Holland describes § 230 as a “modified, less demanding form of cyberlibertarian exceptionalism” that was “generally intended to provide online service providers and bulletin board hosts with immunity from tort liability from the defamatory acts of their users” (at 104). 255 Frank H Easterbrook “Cyberspace and the Law of the Horse” (1996) U Chi Legal F 207 at 207. 256 At 207. 257 Lawrence Lessig “The Law of the Horse: What Cyberlaw Might Teach” (1999) 113 Harv L Rev 501 at 503; see also Lawrence Lessig Code and Other Laws of Cyberspace (Basic Books, New York, 1999). 258 At 511; ‘architecture’ in the sense that some areas of cyberspace require passwords to be accessed, or that some areas of the internet allow encryption but others do not. 259 Lessig was the 9th most cited legal academic in the United States in the period 2009-2013: Brian Leiter “Ten most cited law faculty in the U.S. 2009 through 2013” <http://leiterlawschool.typepad.com/leiter/2014/06/ten-most-cited-law-faculty-in-the-us-from-2009-through-2013.html> accessed 29 September 2015. 260 Lawrence Lessig Code Version 2.0 (2006) <http://codev2.cc/download+remix/> downloaded 29 September 2015.

43      

Mark Tushnet identifies two strategies that courts and legislatures have used to deal with

technological innovations: “one allowing legislative experimentation until experience

accumulates, the other imposing existing doctrine from the outset”.261 The first strategy

implies exceptionalism, at least for a period of time until judges and legislators begin to

understand the new technology.262 However, the second doctrine holds that judges and

legislators accumulate knowledge across technologies, meaning it would be legitimate to

make conclusions about the internet based on existing knowledge of – for example – film,

television and radio.263

Under the first doctrine, where the legislature does not intervene, there is a risk of

disconnection between law and technology. Judges are left to do their best to reason, often by

analogy, from the common law.264 This risk of disconnection has not gone unnoticed in the

cases that apply tortious defamation principles to internet media. Turner notes that Kirby J

was clearly alive to this tension in Gutnick:

“there are a number of difficulties that would have to be ironed out before the settled rules of defamation law … could be modified in respect of publication of allegedly defamatory material on the Internet”.265

But Roger Brownsword points out that the risk of disconnection is inevitable with any new

technology:

“If connection is a generic challenge, we can assume that each technology will be developing at a rate that threatens disconnection. To be sure, some technologies will develop even more quickly than others; but, generally, this will not be where the relevant difference is to be found.”266

His point is a challenging one. The speed with which the internet has transformed so many

aspects of modern life is remarkable, and it is tempting to enlist this fact in support of the

proposition that the internet warrants an exceptional regulatory response. But even if we

cannot, the internet is so exceptionally different to print media that it represents an entirely

different paradigm.267 The cost of publication (and re-publication) has reduced to virtually

                                                                                                               261 Tushnet, above n 86, at 1639. 262 At 1644. 263 At 1645. 264 See 4.4 above on the use of analogies in legal reasoning. 265 Turner, above n 91, at 41; citing Dow Jones & Company Inc v Gutnick, above n 35, at 632. 266 Roger Brownsword and Morag Goodwin Law and the Technologies of the Twenty-First Century (Cambridge University Press, Cambridge, 2012) at 31. 267 See 2.2 above.

44      

nothing.268 Content is easily accessible via sophisticated search engines, and, once published,

remains retrievable indefinitely.269 These features amount to a sea change in the conditions

faced by the defamation tort.

Admittedly, whether this new paradigm of online publication is so different that it warrants

legislative intervention is a different question, one whose answer depends to a large extent on

the form such intervention would take. However, given the authorities thus far constitute a

“growing volume of contradictory precedents”,270 legislative intervention is an attractive

option.

5.3 The Harmful Digital Communications Act: safe harbour provisions

The legislature has in fact already enacted internet-specific rules, in the Harmful Digital

Communications Act (‘the HDCA’). The HDCA was enacted on 2 July 2015 in response to

concerns over cyber-bullying.271 It has a significant potential effect on the third party liability

of Facebook hosts for defamation.

The HDCA created a new offence of posting a digital communication intending to cause

harm.272 However, the HDCA also sets out ‘safe harbour’ provisions, which, if followed,

protect an online content host against most forms of liability for content posted on his or her

page by others.273 This includes protection against liability for defamation.274

An ‘online content host’ is defined in the HDCA as:

“the person who has control over the part of the electronic retrieval system, such as a website or an online application, on which the communication is posted and accessible by the user".275

This will almost certainly include Facebook hosts, as the courts are likely to deem them to

have ‘control’ over their Facebook page. Crucially, Facebook hosts have the power both to

block users and remove content.276 Moreover, the Court of Appeal appeared to accept that a

                                                                                                               268 See 2.2.1 above. 269 See 2.2.2 above. 270 Turner, above n 91, at 52. 271 (30 June 2015) 706 NZPD 4850. 272 Section 22. 273 Sections 23-35. 274 Section 24(1). 275 Section 4. 276 See 2.1.2 above.

45      

person “in Mr Murray’s shoes” (a Facebook host) would have been protected by the

equivalent, very similar, piece of legislation in the United Kingdom.277

If the Facebook host receives a notice of complaint about content on his or her page, and

responds in accordance with s 24 of the HDCA, he or she will be protected against almost all

forms of liability (there are some limited exceptions, but defamation is not amongst them).278

Section 24 requires that upon receiving a valid notice of complaint the host must take all

reasonable steps to contact the author of the content in question within 48 hours.279 If the

author consents, the host must take down the relevant content.280 If the author does not

consent, the host must leave the content online, notify the complainant, and, if the author

consents, provide the complainant with the author’s personal details.281

This bolsters the protection afforded to online hosts by the Court of Appeal’s narrowing of

the legal test for third party publication on Facebook.282 Even where the Facebook host does

have actual knowledge of the defamatory material, he or she will be protected from liability if

he or she follows the procedure provided by s 24. And, importantly, following (or attempting

to follow) the s 24 procedure does not prevent the Facebook host from arguing that he or she

was not a publisher of the content in the first place, as s 23 preserves any defences that might

be otherwise available.283

One advantage of the safe harbour provisions is that they clarify the responsibility of the

author of the publication. Once a complaint has been made to the host, provided the host

follows the s 24 procedure (which he or she must, to rely on the defence), there are two

possible outcomes. The content may be taken down – if the host cannot contact the author,

the author fails to respond to the host, or the author consents to the content’s removal.284

Alternatively, the content may not be removed, but in this case the author will have had to

effectively take responsibility for the content by sending written notice to the host refusing its

                                                                                                               277 Murray v Wishart, above n 9, at [146]; see 5.4 below on s 5 of the Defamation Act 2013 (UK). 278 Sections 24(1) and 25(4). 279 Section 24(2). 280 Section 24(2)(c). 281 Section 24(2)(d). 282 Murray v Wishart, above n 9; see 3.3.1 above. 283 Section 23(3). 284 Sections 24(2)(b), 24(2)(e) and 24(2)(c) respectively.

46      

removal.285 This means that the author is identifiable, and has clearly assumed responsibility

for the publication, such that it makes sense to limit the host’s liability.

5.3.1 Remove first; ask questions later

One problem with the safe harbour provisions is that their structure seems disproportionately

weighted towards the removal of content. Where the host cannot contact the author, or the

author does not reply to a notice of complaint, s 24 requires that the content must still be

removed.286 There is no penalty for a false or malicious complaint, so there is a risk that

complainants could abuse the safe harbour provisions to compel the removal of content for

illegitimate reasons.287 Malicious use of ‘take down’ provisions is not a novel phenomenon.

Urban and Quilter note that § 512 of the American Digital Millennium Copyright Act, which

contains a take down mechanism in relation to alleged copyright infringement, was:

“commonly being used … to create leverage in a competitive marketplace, to protect rights not given by copyright (or perhaps any other law) and to stifle criticism, commentary and fair use.”288

The potential for suppression of legitimate free speech is real, especially in cases where the

material is posted anonymously, making the author difficult to contact.

The Office of the Privacy Commissioner does not seem to share this concern, asserting that

the host only needs to remove material where the complaint comes from someone who is

directly affected by the content:

“It’s not enough for content to be generally offensive or outrageous. It needs to be specifically harmful to the individual who complains about it. This means that any “flooding” would be able to be quickly triaged by the content host, and most of it dismissed out of hand.”289

The wording of the safe harbour provisions does not appear to stipulate such a restriction, and

it is unclear on what basis this statement is made. Instead, it seems the most onerous

                                                                                                               285 Section 24(2)(d). 286 Sections 24(2)(b) and (e). 287 Andrew Easterbrook “Harmful Digital Communications “safe harbour”” (21 August 2015) Auckland District Law Society <www.adls.org.nz/for-the-profession/news-and-opinion/ 2015/8/21/harmful-digital-communications-%E2%80% 9Csafe-harbour%E2%80%9D/>. 288 Jennifer M Urban and Laura Quilter “Efficient Process or “Chilling Effects”? Takedown Notices Under Section 512 of the Digital Millennium Copyright Act” (2006) 22 Santa Clara Computer and High Tech LJ 683 at 687. 289 Sam Grover “Harmful Digital Communications Act: two months in” (31 August 2015) Privacy Commissioner Blog <www.privacy.org.nz/blog/harmful-digital-communications-act-two-months-in>.

47      

requirement in the notice of complaint is to explain why the complainant considers that the

content is unlawful or in breach of the HDCA.290

5.3.2 Conclusions on the HDCA

National MP Scott Simpson noted in a speech during the Harmful Digital Communications

Bill’s Third Reading that one of the Bill’s key aims was to protect young people who did not

have the ability to use “the tools that existed prior to this legislation being enacted, such as

taking defamation actions”.291 While the HDCA will likely succeed in addressing some cases

of online defamation, it is not perfect. By implementing a ‘take-down-first, ask-questions-

later’ approach to internet moderation, the HDCA risks upsetting the delicate balance that the

defamation tort attempts to strike in its measured infringement of the right to freedom of

expression.

However, arguably the defamation tort is not doing a very good job of striking this balance in

practice. Defamation proceedings are expensive and lengthy, notoriously so.292 Perhaps the

single most attractive feature of the HDCA safe harbour provisions is that they provide a far

cheaper and simpler avenue of recourse than the traditional litigation route. In being cheaper

and simpler, the safe harbour provisions should make recourse against a tortfeasor more

accessible. Given the greater accessibility and transmissibility of internet communications,

this is both appropriate and desirable.293

5.4 The equivalent scheme in the United Kingdom

Section 5 of the Defamation Act 2013 (UK) provides a very similar ‘take down’ scheme to

the HDCA safe harbour provisions.294 The s 5 scheme provides ‘operators’ of websites with a

complete defence in respect of third party content on their websites if they respond to a notice

                                                                                                               290 Section 24(3). 291 (30 June 2015) 706 NZPD 4850. 292 See, for example, Davis v Independent Newspapers Ltd HC Auckland CP67SD02, 12 March 2003 at [35]: “the process of defending a claim for defamation can be a protracted and expensive exercise”; Awa v Independent News Auckland Ltd [1996] 2 NZLR 184 at 187 per Hammond J: “Defamation proceedings are notoriously expensive”; Police v W [1989] 3 NZLR 696 at 701 per Fisher J: “A civil action for defamation is extraordinarily protracted and expensive compared with virtually all other classes of civil litigation”. 293 See 2.2 above. 294 Defamation Act 2013 (UK), s 5.

48      

of complaint and comply with the take down procedure prescribed by the Act’s

regulations.295

This part of the Act reflects a concern that defamation laws were “not well suited to dealing

with the internet and modern technology”, were “becoming out of date, costly and over-

complicated”, and that there was a “risk of damaging freedom of speech without affording

proper protection”.296

The Joint Committee on the Draft Defamation Bill identified two key problems with the

English law as it stood prior to the Act. The first was the risk aversion of service providers.

Given service providers risked attracting primary liability for content once they received a

complaint, their reaction was often to take down the allegedly offending content whenever a

complaint was received.297 Service providers typically do not have a strong understanding of

defamation law, resulting in the removal of many entirely legitimate comments.298 Rather

than up-skilling, or employing lawyers, to place themselves in a position to make the kind of

judgments necessary to assess liability, service providers were more likely to simply remove

content following a complaint. A second problem was that if service providers did employ

moderators or a moderating system this could actually lead to increased liability. Because

liability was founded on knowledge of the allegedly defamatory content, employing

moderators to actively monitor content meant that providers risked attracting liability earlier

than they otherwise would have, as the moderators became aware of the content.299 Thus:

“As the law stands, far from encouraging service providers to foster legitimate debate in a responsible manner and removing the most extreme material, it encourages them to ignore any dubious material but then to remove it without question following a complaint.”300

The structure of incentives created by the law did not align with the protection of reputation,

nor did it encourage the right to freedom of expression.

                                                                                                               295 Section 5, along with the Defamation (Operators of Websites) Regulations 2013 (UK); see also Ter Kah Leng “Internet defamation and the online intermediary” (2015) 31 CLSR 68 at [2]. 296 (12 June 2012) 546 GBPD HC 177. 297 Joint Committee on the Draft Defamation Bill (UK) Draft Defamation Bill (19 October 2011) at 53-54. 298 At 53-54. 299 At 54. 300 At 54.

49      

The s 5 scheme in the Defamation Act 2013 (UK) is structured in a similar manner to the

HDCA safe harbour provisions. ‘Operator’ will likely have a comparable meaning to ‘online

content host’ in the HDCA.301 Notably, where the operator cannot contact the author of the

content, the author does not reply to the operator’s notice of complaint, or the author sends an

incomplete reply to the operator, the same presumption that the content will be taken down

applies.302 Where the originator responds in accordance with the scheme, the operator must

not remove the content and the operator has a defence to liability for defamation.303 One

difference between the two schemes is that in the original complaint to the operator the

complainant must explain why the statement is defamatory of the complainant.304 However,

since the Act does not require the operator to verify this claim, a mere assertion might

suffice, meaning the s 5 scheme is potentially exposed to malicious complaints in the same

manner as the HDCA.305

One of the key attractions of the s 5 scheme is that, like the HDCA safe harbour provisions, it

clarifies the responsibility of the author. Turner lauds the s 5 scheme for facilitating

conversations:

“between the claimant and the [author] (the latter having full knowledge of the content and context of the material and who is, therefore, in the best position to determine whether the content ought to be removed)”.306

As explored above, this will likely achieve a clearer delineation of responsibility for the

publication, by compelling the author to either take responsibility for the content (in writing),

or risk having it taken down.307

The second key benefit of the HDCA safe harbour provisions also applies to the s 5 scheme:

it allows for recourse against distributors of defamatory third party content outside of actual

                                                                                                               301 See 5.3 above; ‘operator’ is defined, in para 1(2) of the Defamation (Operators of Websites) Regulations 2013 (UK), as “the operator of the website on which the statement complained of in the notice of complaint is posted”, but this does little to clarify the scope of the definition. 302 Defamation (Operators of Websites) Regulations 2013 (UK), paras 2(3), 5 and 6 of the Schedule respectively; see 5.3.1 above. 303 Defamation (Operators of Websites) Regulations 2013 (UK), para 8 of the Schedule. 304 Defamation Act 2013 (UK), s 5(6)(b). The equivalent requirement under New Zealand’s Harmful Digital Communications Act 2015, found in s 24(3), is much broader, requiring the complainant to set out why he or she considers the content unlawful or in breach of the HDCA; see 5.3.1 above. 305 See 5.3.1 above. 306 Turner, above n 91, at 62. 307 At 5.3.

50      

or threatened defamation litigation.308 The time and expense associated with bringing a

proceeding in defamation makes the existence of a simpler and cheaper avenue of recourse

against a tortfeasor hugely attractive.309

5.5 Conclusion

The nascent safe harbour provisions in the HDCA have the potential to greatly affect the third

party liability of Facebook hosts in New Zealand. Indeed, the Court of Appeal in Murray v

Wishart appeared to accept that a person “in Mr Murray’s shoes” would be protected by s 5

of the United Kingdom’s Defamation Act.310 Presumably, then, given the similarities

between s 5 and the safe harbour provisions of the HDCA, Mr Murray would have been able

to use the safe harbour provisions to his benefit. The HDCA has therefore bolstered the

protection afforded to online hosts already strengthened by the Court of Appeal’s narrowing

of the legal test for third party publication on Facebook. While it does have the potential to be

abused through malicious use of the safe harbour provisions, it also has great potential as a

cheaper and more direct method of protecting the rights to speech and reputation online.  

                                                                                                               308 Turner, above n 91, at 62; see 5.3.2 above. 309 See 5.3.2 above. 310 Murray v Wishart, above n 9, at [146].

51      

CONCLUSION

In applying the defamation tort to the internet, the fundamental challenge is in working out

how the tort can continue to effectively protect reputational interests without undermining the

great potential of the internet as a medium for public expression. However, publication on the

internet is significantly different to publication via traditional print media. Compared to print

media, publication on the internet is characterised by the ease with which material can be

published, the increase in the material’s longevity, the ease with which material can be

retrieved, and the ability to publish anonymously. The cost of publishing something on the

internet has reduced to virtually zero, meaning the potential for republication is effectively

unlimited. This new paradigm of publication represents a set of challenging conditions for the

defamation tort. Indeed, these changes to the way society communicates have rightly been

called “revolutionary”.311

That the courts have had difficulty in applying the defamation tort to internet media is

therefore unsurprising. In the Murray v Wishart proceedings the High Court and the Court of

Appeal had to decide how to determine the liability of a Facebook host for comments posted

on his page by third parties. Both courts applied Byrne v Deane publication by omission

principles to assess the Facebook host’s liability for third party content. The Hong Kong

Court of Final Appeal took a different approach in a case that had some similar features,

using publication by action principles to assess the liability of forum providers for third party

content.

Neither of these approaches is perfect. While the publication by action approach would hold a

Facebook host to a higher level of liability, given the statutory construction of the innocent

dissemination defence in New Zealand, it would likely lead to an undesirably broad

application of liability.312

It seems then that the Byrne v Deane line of authority must provide the appropriate legal

framework. The High Court and Court of Appeal differed as to the type of knowledge

required to found publication by omission. While constructive knowledge allows for a more

                                                                                                               311 Solove, above n 90, at 17. 312 See 1.4.1 and 4.3.2(b) above.

52      

nuanced application of Byrne v Deane principles, what it gains in flexibility it sacrifices in

certainty. Conversely, while the actual knowledge test provides greater certainty, its rigidity

means that the circumstances surrounding the Facebook page cannot affect the assessment of

liability.

Ultimately the question of what is the appropriate legal test, and the difficulties therein, gives

way to a larger question. If the attempt to apply the principles of defamation law to internet

media is inevitably going to be clumsy, could the better course be to implement technology-

specific rules governing defamation-type situations online?

That is what the enactment of the Harmful Digital Communications Act 2015 hopes to

achieve. The safe harbour provisions in the HDCA have the potential to greatly affect the

third party liability of Facebook hosts in New Zealand. They have bolstered the protection

afforded to online hosts already strengthened by the Court of Appeal’s narrowing of the legal

test for third party publication on Facebook. A Facebook host who follows the procedure

prescribed in the HDCA will be protected from liability for defamation. While the HDCA

does have the potential to be abused through malicious use of the safe harbour provisions, it

also has great potential as a cheaper and more direct method of protecting the rights to speech

and reputation online. For this reason it should be welcomed.

As the internet continues to transform the modern world, it will continue to challenge

established common law concepts, and make demands on those responsible for the upkeep

and development of the law. Defamation’s intersection with the internet ultimately

demonstrates the difficulty that the common law, and the defamation tort in particular, has in

coping with the upheavals of the internet age.

53      

BIBLIOGRAPHY

 A. CASES 1. New Zealand

Awa v Independent News Auckland Ltd [1996] 2 NZLR 184.

Bell-Booth Group Ltd v Attorney-General [1989] 3 NZLR 148 (CA).

Brooker v Police [2007] NZSC 30.

Davis v Independent Newspapers Ltd HC Auckland CP67SD02, 12 March 2003.

Living World Distributors Ltd v Human Rights Action Group Inc (Wellington) [2000] 3 NZLR 570; (2000) 6 HRNZ 28 (CA).

Murray v Wishart [2014] NZCA 461.

Police v W [1989] 3 NZLR 696.

Sadiq v Baycorp (NZ) Ltd HC Auckland CIV-2007-404-6421, 31 March 2008.

Television New Zealand Ltd v Keith [1994] 2 NZLR 84 (CA).

Wishart v Murray [2013] NZHC 540.

2. United Kingdom

Bunt v Tilley & Ors [2006] EWHC 407 (QB).

Byrne v Deane [1937] 1 KB 818.

Cairns v Modi [2013] 1 WLR 1015; [2012] EWCA Civ 1382.

Davison v Habeeb [2011] EWHC 3031 (QB).

Emmens v Pottle (1885) 16 QBD 354.

Godfrey v Demon Internet Ltd [2001] QB 201.

Handyside v United Kingdom (1976) 1 EHRR 737.

Home Office v Harman [1983] 1 AC 280.

Metropolitan International Schools Ltd v Designtechnica Corporation [2009] EWHC 1765 (QB); [2011] 1 WLR 1743.

Monson v Tussauds Ltd [1894] 1 QB 671.

Panday v Gordon [2006] 1 AC 427.

Parmiter v Coupland (1840) 6 M&W 105.

Pullman v Walter Hill & Co [1891] 1 QB 524.

Scott v Sampson (1882) 8 QBD 491.

Sim v Stretch (1936) 52 TLR 669 (HL).

Tamiz v Google Inc [2012] EWHC 449 (QB); [2013] EWCA Civ 68.

Vizetelly v Mudie’s Select Library Limited [1900] 2 QB 170.

Youssoupoff v Metro-Goldwyn-Mayer Pictures Ltd (1934) 50 TLR 581 (CA).

54      

3. Australia

Dow Jones & Co Inc v Gutnick [2002] HCA 56; (2002) 210 CLR 575.

Frawley v State of New South Wales [2006] NSWSC 248.

Thompson v Australia Capital TV Ltd (1996) 186 CLR 574.

Urbanchich v Drummoyne Municipal Council (1991) Aust Tort Reports 81-127 (NSW SC).

4. Canada

Crookes v Newton [2011] 3 SCR 269.

Hill v Church of Scientology of Toronto [1995] 2 SCR 1130.

R v Salituro [1991] 3 SCR 654.

5. Hong Kong

Oriental Press Group Ltd v Fevaworks Solutions Ltd [2013] HKCFA 47.

6. United States of America

Cubby Inc v Compuserve Inc 776 F Supp 135 (SD NY 1991).

Isabelle Hellar v Joe Bianco (1952) 244 P 2d 757.

Ostrowe v Lee 256 NY 36 (NY Ct App 1931).

Rosenblatt v Baer 383 US 75 (1966).

Scott v Hull 259 NE 2d 160 (Ohio Ct App 1970).

Stratton Oakmont Inc v Prodigy Services Co 1995 WL 323710 (NY Sup Ct 1995).

Tacket v General Motors Corp 836 F 2d 1042 (7th Civ 1987).

7. European Union

Case C-131/12 Google Spain SL and Google Inc v Agencia Española de Protección de Datos (AEPD) and Mario Costeja González OJ 2014/C 212/04.

B. LEGISLATION 1. New Zealand

Defamation Act 1992.

Harmful Digital Communications Act 2015.

New Zealand Bill of Rights Act 1990.

2. United Kingdom

Defamation Act 2013.

Defamation (Operators of Websites) Regulations 2013.

3. United States of America

Communications Decency Act 1996.

55      

C. BOOKS Roger Brownsword and Morag Goodwin Law and the Technologies of the Twenty-First Century (Cambridge University Press, Cambridge, 2012).

Peter F Carter-Ruck, Richard Walker and Harvey NA Starte Carter-Ruck on Libel and Slander (4th ed, Butterworths, London, 1992).

Matthew Collins The Law of Defamation and the Internet (3rd ed, Oxford University Press, Oxford, 2010).

William Holdsworth History of English Law (7th ed, Methuen, London, 1972).

Michael A Jones, Anthony M Dugdale and Mark Simpson Clerk & Lindsell on Torts (21st ed, Sweet & Maxwell, London, 2014).

Lawrence Lessig Code and Other Laws of Cyberspace (Basic Books, New York, 1999).

EH Levi An Introduction To Legal Reasoning (University of Chicago Press, Chicago, 1948).

Patrick Milmo and WVH Rogers Gatley on Libel and Slander (10th ed, Sweet & Maxwell, London, 2004).

Brian Neill, Richard Rampton, Heather Rogers, Timothy Atkinson and Aidan Eardley Duncan and Neill on Defamation (3rd ed, LexisNexis, London, 2009).

Friedrich Nietzsche The Gay Science (Cambridge University Press, Cambridge, 2001).

Geoffrey Palmer Constitutional Conversations: Geoffrey Palmer talks to Kim Hill on National Radio (Victoria University Press, Wellington, 2002).

David Rolph Reputation, Celebrity and Defamation Law (Ashgate, Aldershot, 2008).

Daniel J Solove The Future of Reputation: Gossip, Rumor, and Privacy on the Internet (Yale University Press, New Haven, 2007).

D. JOURNAL ARTICLES Jacqueline Bryant “Defamation Act 1992” (1993) 7 Auckland U L Rev 417.

Ursula Cheer “Defamation in New Zealand and Its Effects on the Media – Self-Censorship or Occupational Hazard?” (2006) NZ L Rev 467.

Joachim Dietrich “Clarifying the meaning of ‘publication’ of defamatory matter in the age of the internet” (2013) 18 Media and Arts Law Review 88.

Frank H Easterbrook “Cyberspace and the Law of the Horse” (1996) U Chi Legal F 207.

Eleni Frantziou “Further Developments in the Right to be Forgotten: The European Court of Justice’s Judgment in Case C-131/12, Google Spain, SL, Google Inc v Agencia Española de Protección de Datos” (2014) 14 HRL Rev 761.

H Brian Holland “In Defense of Online Intermediary Immunity: Facilitating Communities of Modified Exceptionalism” (2007) 56 U Kan L Rev 369.

Andreas M Kaplan and Michael Haenlein “Users of the world, unite! The challenges and opportunities of social media” (2010) 53 Business Horizons 59.

Grant Lamond “Analogical Reasoning in the Common Law” (2014) 34 Oxford Journal of Legal Studies 567.

Ter Kah Leng “Internet defamation and the online intermediary” (2015) 31 CLSR 68.

Lawrence Lessig “The Law of the Horse: What Cyberlaw Might Teach” (1999) 113 Harv L Rev 501.

Lyrissa Barnett Lidsky “Silencing John Doe: Defamation & Discourse in Cyberspace” (2000) 49 Duke LJ 855.

Angus Macinnis “Guilt written all over your Facebook: Liability for publication in social media” (2011) 49 LSJ 30.

56      

Richard A Posner “Book Review: Reasoning By Analogy” (2006) 91 Cornell L Rev 761.

David Rolph “Defamation by social media” (2013) 117 Precedent 16.

Jonathan Scragg and Caitlin Ashworth “Like and Share: Publisher Liability for Third-Party Posts on Facebook” (2014) 8 NewLaw 10.

Ryan J Turner “Internet Defamation Law and Publication by Omission: A Multi-Jurisdictional Analysis” (2014) 37 UNSWLJ 34.

Mark Tushnet “Internet Exceptionalism: An Overview From General Constitutional Law” (2015) 56 Wm and Mary L Rev 1637.

Jennifer M Urban and Laura Quilter “Efficient Process or “Chilling Effects”? Takedown Notices Under Section 512 of the Digital Millennium Copyright Act” (2006) 22 Santa Clara Computer and High Tech LJ 683.

E. PARLIAMENTARY MATERIALS 1. New Zealand

(30 June 2015) 706 NZPD 4850. 2. United Kingdom

(12 June 2012) 546 GBPD HC 177.

F. REPORTS 1. New Zealand

Law Commission Electronic Commerce Part 2: A basic legal framework (NZLC: R58, 1999). 2. United Kingdom

Joint Committee on the Draft Defamation Bill Draft Defamation Bill (19 October 2011).

G. SPEECHES Robert Ribeiro “Defamation on the Internet” (paper presented to Obligations VII Conference, Hong Kong, July 2014).

Jack Balkin “Media Access: A Question of Design” (paper presented to the Global Constitutionalism Seminar, Yale Law School, 2010).

H. INTERNET RESOURCES John Perry Barlow “A Declaration of the Independence of Cyberspace” <https://projects.eff.org/~barlow/Declaration-Final.html>.

Andrew Easterbrook “Harmful Digital Communications “safe harbour”” (21 August 2015) Auckland District Law Society <www.adls.org.nz/for-the-profession/news-and-opinion/ 2015/8/21/harmful-digital-communications-%E2%80% 9Csafe-harbour%E2%80%9D/>.

Sam Grover “Harmful Digital Communications Act: two months in” (31 August 2015) Privacy Commissioner Blog <www.privacy.org.nz/blog/harmful-digital-communications-act-two-months-in>.

Brian Leiter “Ten most cited law faculty in the U.S. 2009 through 2013” <http://leiterlawschool.typepad.com/leiter/2014/06/ten-most-cited-law-faculty-in-the-us-from-2009-through-2013.html>.

Lawrence Lessig Code Version 2.0 (2006) <http://codev2.cc/download+remix/>.

57      

Jon Ronson “How One Stupid Tweet Blew Up Justine Sacco’s Life” The New York Times Magazine (online ed, New York, 15 February 2015).

Facebook Help Center <www.facebook.com/help>.

Facebook Newsroom <http://newsroom.fb.com/company-info>.

<www.firstdigital.co.nz/blog/2014/09/10/facebook-demographics-new-zealand-age-distribution-gender>.

<www.huffingtonpost.com/brian-honigman/100-fascinating-social-me_b_2185281.html>.

<www.huffingtonpost.com/jo-brothers/a-social-media-breakfast-_b_2703179.html>.